FiscalNote Holdings, Inc. (NOTE) Earnings Call Transcript & Summary

First, we'll hear from FiscalNote's Chairman, CEO and Co-Founder, Tim Hwang.

Good morning, everybody. I want to welcome everyone to FiscalNote's 2024 AI Product Day. My name is Tim Hwang, Chairman, CEO and Co-Founder of FiscalNote. What I want to do today before we start is talk a little bit about FiscalNote, how we got started, how we were founded, the mission of the company, and of course, walk through what we hope to accomplish here over the course of the next several hours. FiscalNote, we started this company 11 years ago with very humble beginnings. And we started off with the mission of wanting to connect the world to their governments. We were just 3 young guys in our early 20s just graduating college. And we had this very simple idea to essentially build an information company that aggregated legislation, regulations, court cases, government filings from as many different countries around the world as possible, and to be able to take that information and help customers, companies, law firms, government agencies, really understand how the world is changing and how regulations may potentially impact their particular organizations. We started off working out at my parent's basement, really trying to code the product ourselves, talk to customers and really build a long-lasting product that can be used by as many people as possible. From there, we went off and we bought a one-way ticket to Silicon Valley, and we were living at a Motel 6 room, really just working day-to-day, week-to-week, trying to build as best a product as possible. I think we're probably paying ourselves very, very little. And I think that the foundations of FiscalNote really started with the mission of the company of wanting to really get high-quality legal, regulatory and political information into the hands of our customers. I remember one evening, I think it was in late 2013, we were watching Shark Tank after a really long night, and we turned to ourselves and said, "Hey, wouldn't it be really cool if Mark Cuban invest in our company?" And so went on Google, typed in Mark Cuban's e-mail address, shot him a cold e-mail and then a couple of days later, Mark ended up being our first investor. For Mark, to Jerry Yang, to Steve Case, to NEA, to various other top-tier venture investors, we were able to build our company from scratch. And we went off and we moved the company's headquarters from Silicon Valley to D.C., really setting up our first office, no heating, no air conditioning, no windows, with borrowed furniture from Craigslist. And we are growing the business step by step, product by product, person by person, year after year. And it's been an incredible journey here at FiscalNote, really trying to make a name for ourselves in our space. One of the things I'll point out is that we were an AI company 11 years ago before AI became cool. And a lot of what we're trying to do from the very beginning was thinking about how we could apply artificial intelligence to helping our customers really understand how laws and regulations may potentially impact their institutions. And so if you turn to the newspapers from 2013, 2014, you'll probably recall headlines like Uber executives or Airbnb executives or Zenefits executives or whatever the case may be, really wrestling with regulatory issues from 10-plus years ago. And really from that regulatory environment, FiscalNote was born and trying to help many, many companies since then really understand regulations and the impacts on their industries. So since then, we've expanded our company tremendously. We service thousands of customers around the world, ranging from hundreds of government agencies, ranging from the White House to Congress to the Federal Reserve, over half the Fortune 100 currently use FiscalNote today in their legal, regulatory, public policy or government affairs offices. And many, many organizations on -- in [ casuistry ] effectively are using FiscalNote to really understand how laws and regulations are moving through our democratic processes. Now over the course of the last several years, what we tried to do is really think about how to apply cutting-edge technologies and really understanding those laws and regulations. And so over the last several months, we've made a significant level of investments in artificial intelligence and particularly really leaning into the new capabilities of generative AI. We laid out very extensively our AI strategy and what we're investing in for the future. And that AI strategy starts, first and foremost, from the foundation of high-quality information. Now I've said for the last 10, 11 years, we've been aggregating legislation, regulations, policy updates, whatever the case may be, around 80 different countries, everything ranging from environmental regulations from Argentina to real estate regulations in Indonesia, cybersecurity regulations in Australia, effectively trying to get a really good understanding of high-quality information for laws and regulations. Now in order to do that, we've used quite extensive information kind of enhancing tools, things like machine translation or things like automated summarization or other capabilities that we build proprietary technologies on. Now from there, what we've done is we've started to think about what's next for the company. We are already servicing all these customers with high-quality information. And so of course, the next step was really helping our customers to understand what they should do with that information. And so over the course of the last several quarters, we, for instance, took the first step by partnering with OpenAI, Microsoft and Google in their generative AI efforts in a variety of contexts. Then what we did was we started building our own generative AI tools internally. We built our own reasoning engine, for instance, that fact checks these language models to understand whether or not what's coming out of kind of generative AI is true or not based purely off the information that we have. We built an AI Copilot Creator that effectively enables us to able to build our own copilots and our own chatbots using the information that we have. And so today, many of our leading technologists, data scientists and executives are effectively going to walk you through many of those investments and really what's in store for FiscalNote in the future. And so with that, I want to talk a little bit about sort of where we're headed and of course, some of the agenda for today. Now based purely off the information that we have today, the AI investments that we've made, we're now embarking on the strategy of launching a variety of new generative AI products into the marketplace, a variety of new skills and capabilities and copilots that effectively help our customers to extend the information that we have into their everyday workflows. And so we're going to start off the day with our President and COO, Josh Resnik, who's really going to provide an overview of our current products and services. He's going to explain how our data impacts important decisions around the world and why it matters. Then of course, we've got our CTO and our Chief Scientist, Vlad Eidelman, who will speak about our innovation and the new products that we've been pioneering in the use of machine learning and other AI techniques and policies. Next, we'll go over our Copilot for Global Intelligence, which really accelerate the discovery of relevant content and create an experience that connects surfaces insights from a variety of different products. Then we're going to dive into our FiscalNote Risk Connector, which we've launched in response to customer demands for better, faster and more data-driven ways to map their supply chains and identify risk far in advance of when they materialize. Next, I think after that, we're going to have a really deep discussion around StressLens, which quantifies the human elements of stress and emotion and public statements by policymakers and business leaders, which no one really does. We really leverage AI and data science to detect signals not just in what leaders are saying, but also how they feel about it. Then we'll discuss how we're utilizing the capabilities from Roll Call, Factba.se, StressLens with deep analysis that Factba.se and StressLens products bring to the table. Lastly, we'll introduce Copilot for Policy, our newest AI tool before going into Q&A. We have an incredibly exciting day here today filled with new product enhancements, new product announcements. We've got an incredible lineup of our executives here who spent the last several months and years building an incredible lineup of AI products we're bringing to market. And so we're very excited to bring these products to you today. Thank you for joining us.

Next, we'll hear from FiscalNote's President and Chief Operating Officer, Josh Resnik, who will be going over current products and services.

Thank you all for joining us today. Along with Tim, it's my pleasure to welcome you to this exciting presentation where we dive into some of FiscalNote's newer innovations designed to help you anticipate, interact with and respond to change. Whether you're a leader of a global enterprise or advocating on a local policy issue, we understand that the decisions you are making daily can feel monumental. We're here to help. Having accurate, timely and usable information and being well positioned to act on it is crucial. Our products and services, including our advanced AI tools and unmatched data, empower our customers to make informed decisions with confidence. We're trusted by the world's largest corporations, the most influential government organizations, and the leading associations and NGOs that shape policy and regulations around the world. Though we're showcasing a sample of our technology innovations today, we don't stop at technology. At FiscalNote, we believe that the best results come from augmented intelligence, combining AI with human intellect. Only FiscalNote integrates advanced AI capabilities with the expertise of skilled, seasoned and objective analysts, researchers and journalists. This unique synergy allows us to deliver intelligence that goes beyond raw data. Through our research and customer reports, we ensure our customers receive insights that are not only accurate, but also deeply relevant to their specific needs. Why do our offerings matter? Because the pressure of making a big decision under a cloud of uncertainty is overwhelming for anyone. At FiscalNote, we provide our customers with clarity and foresight to manage geopolitical, economic, security and operational risk. This enables corporations to protect their supply chains, select the right location for their new regional hub, equip their lobbyists with the most accurate data and deploy their advocates with the right messages. It's important to note that this creates a critical downstream effect because when we supply governments, associations and corporations with more intelligence, they make higher quality and more beneficial decisions for the individuals they serve, be it constituents, consumers or advocates. That's why our purpose at FiscalNote is to create a more transparent and informed global society. In a rapidly evolving world, the value of AI lies in its unparalleled ability to synthesize vast amounts of data and content into precise real-time insights. As AI continues to evolve, it will become an indispensable tool for negotiating the complexities of international policy and market dynamics. Our customers will now be able to anticipate and respond to emerging threats and opportunities with unprecedented speed and accuracy. Ultimately, AI will not only enhance our understanding of global issues, but also empower us to shape a more secure and prosperous future. To our valued customers, I want to especially emphasize that the AI innovations you see today are not just about checking an AI box or being part of a trend. We're leveraging the best in technology to drive meaningful value, providing you with the support you need to be successful. You can expect that many of the technological innovations you see today will be integrated into our core products in the coming months. These advancements are meaningful because they save you time, empower you to optimize and better manage critical resources, enhance your strategic planning and provide you with a competitive edge. By combining advanced AI with human expertise, we ensure that you receive the most accurate and relevant intelligence, helping you stay ahead of the curve and effectively respond to emerging challenges and opportunities. Our commitment to leveraging AI means that you can focus on what matters most, achieving your goals and driving success in an ever-changing world. Thank you for trusting FiscalNote to be your partner. We're dedicated to delivering the best solutions to empower your decision-making and strategic initiatives. And thanks to everyone for joining us today. We look forward to continuing this journey of innovation and excellence with all of you.

Next up is FiscalNote's Chief Technology Officer and Chief Scientist, Vlad Eidelman, who will be speaking about innovation and new products.

Good morning. I'm Vlad Eidelman, Chief Technology Officer. I joined FiscalNote over 10 years ago as one of our first 10 employees and built out the first versions of many of our AI-enabled features, and I'm now responsible for leading the global technology team. Before we dive into the really exciting demos of what we've been working on recently, let me set the stage for how these new products fit naturally into the broader arc of our last decade of innovation that Tim and Josh have just been talking about. So for years, we've been pioneering the use of machine learning and other AI techniques in this policy space. So for example, we were among the first in the market with machine-learning models that predict the likelihood of legislative success and stakeholder support for every piece of legislation across the U.S. So, customers can prioritize the time on initiatives and people that matter most. We built the first versions of automatically identifying similar policy initiatives across different localities and time periods. We mapped how pending legislation will impact existing laws, identified topics and related policies from local levels up to improve finding that needle in the haystack, and we build models that analyzed every publicly available comment for the stance on proposed regulations so customers can get a sense of where things might go. So why do we build all of these or better yet, what are customers really looking for from any of our products? You see it's not about wanting AI for the sake of AI. We invest in AI because our customers need accurate data fast. And even more than that, we invest in AI because equipped with the right information, our customers can change the world to deliver the insights and make the connections that our customers rely on. AI has been at the heart of our strategy since 2013. We really couldn't handle the tens of thousands of data sources and process the millions of documents each week without it. These documents come from all over the place, whether it's parliamentary proceedings, meeting memos from municipal governments, news articles, financial disclosures, earnings calls, it would be impossible to keep up. Our AI and machine learning models do a lot of the heavy lifting, which helps us serve thousands of customers effectively and efficiently. However, the complexities in the political, the corporate, the regulatory areas sometimes need deep expertise in technology and law. That's why our internal advisory, editorial and analysts play such an important role. Because of that role, we see AI as a tool for augmented intelligence. We believe AI should work alongside human intelligence, enhancing and supporting it. Our AI systems rely on trusted public data and in-house analysis that we've done. And where it makes sense, we have a human review before something goes out to our customers. So take our summarization model, for example. It produces legal summaries for thousands of bills each day or our automatic speech recognition and speaker labeling models, which create near-live earning calls or professional transcripts. These all go directly into our products. But if the quality isn't there, we can route them to a human analyst for a final review. By integrating AI into our human-driven processes, we automate where it makes sense, offering our services that we couldn't before at lower costs and faster speeds, while still maintaining the high quality our customers depend on. So whether it's understanding school board issues in the City Council in Springfield, congressional hearings on AI, changing environmental standards in the EU, our customers need to understand what the primary documents and data are about, why it's important, who is involved and how it affects them, and last, what comes next. That's why we invest in AI. I won't spend much time diving into the automated ingestion engines and tooling that we've built and really honed to manage the vast amounts of data from numerous sources, but that's the foundation for all the varied experiences that you're about to see. They're valuable precisely because we make sure our customers can trust the completeness, the integrity, consistency and the timeliness of our data, which is so crucial for all of those who rely on receiving timely alerts. We've built smarter data ingestion tools that can recognize different data sources and different types and adapt to changes in the missing data. We've also been automatically extracting and summarizing data with advanced ML models and techniques way before the advent of LLMs. We've put a lot of effort into combining high-speed collection and validation capable of synthesizing raw information and standardizing it with the automated data monitoring and observability systems that feed a set of internal tools where human expertise really refines the data throughout the pipeline. This allows us to do most of the work automatically while still maintaining high data quality and providing feedback to improve the system. While the majority of our collection is text, it's a good time to point out that we're actually increasingly becoming multimodal. We're processing more video and more audio streams as well. In any form, while on the surface, that data can be a dense policy or economic discussion inside that regulatory comment from a telecom in response to a new SEC proposed rule are explicit reasons why and how it's going to affect them financially. And inside the farm bill are huge cost and subsidy proposals that will directly affect billions in spending. Inside earnings reports are policy risk the company is already looking at. So from all of this unstructured data, we've been extracting elements of structure and deriving new intelligence automatically for years. More recently, we've expanded a number of offerings with AI capabilities. In our Grassroots advocacy solution, we integrated LLM capabilities to offer a better experience in drafting parts of an e-mail communication, so customers can really ensure the optimal message reaches the audience. For our EU policy analysis solutions and our state and federal government constituent management solutions, we expanded our transcription and analysis offerings to provide the transcription of video and audio streams at the EU and state level hearings so folks can get fast notifications if something relevant is being discussed without even having to monitor everything themselves. It's actually really cool that in developing all of these ML and AI models, we've also built many proprietary training sets from generic entities, like organizations and people, to more specific legal entities, to multiple topic and theme taxonomies that cover thousands of policy and related topics, to industry classifications and risk categories and indicators, to finally policy and geopolitical impact ratings. Conceptually, these are all different ways of automatically deriving really unique metadata and form several knowledge graphs rich with entities and relationships that significantly increase the compounded value of our data. So in the graphs, you can think of nodes representing entities like policy docs, news, regulators and companies while the relationships between the nodes are the enrichments that our machine learning models have made. For example, if you're interested in data privacy around the world, these models could have tagged related news, policies and companies, identified key policymakers and stakeholders and analyzed who's been most effective in pushing these policies and created relationships to concepts like cybersecurity or GDPR for further exploration. Our knowledge graphs allow us to uncover valuable connections among organizations, people, documents, events, enhancing knowledge discovery. And as the graph grows, it becomes even more powerful incorporating more entities and more relationship types. This growth allows us to create various aggregations and perspectives, enabling us to derive actionable insights tailored to specific existing products or to expand into new areas for our customers quickly. So with that, thanks for listening, and I hope you enjoy the demos that we've prepared for you coming up next.

Next, we have Josh Haecker, Head of Product for Global Intelligence; and Rebecca Palser, Chief Content Officer, who would be going over Copilot for Global Intelligence.

Hi. I'm Rebecca and I manage the brilliant teams of analysts across FiscalNote Global Intelligence. We help our customers to understand the macroeconomics, security and geopolitical issues so that they can manage their businesses effectively. Our work enables organizations to be on the front foot as events occur. We provide intelligence and analysis to help them make informed decisions about which markets they should be entering, deciding if it's safe to continue operations in an area near a conflict zone or understanding long-term geopolitical trends that will directly impact on their company's strategy. We help in all of these scenarios and more. Our analysis doesn't just tell them what is happening today. To help our customers stay ahead, our intelligence provides an early warning system by fusing human expertise, knowledge, data and technology. Our customers trust us to tell them what they need to know. Here's what one of them had to say. You are the provider we use the most. As a naval intelligence professional of 20 years, it is the only vendor service I have seen that is actually what I'd call intelligence. However, our role at FiscalNote Global Intelligence is not just to provide expert analysis, but crucially to cut through the noise to enable customers to respond quickly. Our customers do not want to become an analyst themselves. They trust us to tell them what they need to know at the right time. We provide them with an understanding of what's relevant, how things have changed and present a high-level overview of the situation. Our reports provide customers with valuable archives of analysis, but as our library has grown over time, we need to ensure that they can still access the most up-to-date and accurate information with ease, giving them the confidence to act quickly. To make sure that we continue to respond to our customers' need for this level of accessibility, we believe in providing them with the right tools and technology.

That's right, Rebecca. I'm Josh Haecker, the Head of Product for Global Intelligence. And my team has developed the Copilot for Global Intelligence that we'll be talking about in this segment. But first, a little bit about why we felt a copilot was the right tool for our customers. We always approach customer challenges guided not by what we can do, but guided by what delivers the most value. At their core, the main 2 problems we distilled from this feedback were: first, there is so much content, it's hard to find the most relevant data to their specific questions; second, many of our customers subscribe to multiple products across the global intelligence offerings. And while we've received positive feedback on the value of each, the need to move between those products was preventing customers from taking advantage of them fully as an integrated offering. So to deliver the most value, we needed to accelerate discovery of relevant content and create an experience that broke down the barriers between our different products. By combining FiscalNote's long history of innovating in AI alongside the decades of experience from our global intelligence experts and markets and geopolitics, we've been able to develop a product that significantly improves the value we bring to our customers. Following successful beta, we're confident this Copilot returns highly relevant and timely summaries of content across all FNGI products with an emphasis on explicitly citing and linking to the human analysis driving its answers. This allows our users to have confidence that our analytic content is as high quality as ever and allows direct access to the deeper analysis, if needed, regardless of which product it resides in. That's enough for me, though, let's see it in action. Rebecca, why don't we go ahead and ask a few questions in the Copilot for Global Intelligence.So let's just start off with a softball here. What's next for Russia and Ukraine? So right now, as you'll see on the screen, it's digging through those decades of content experience and expert analysis to figure out what's the right answer to the question, and the right answer to the question right now. Once it's done that and tabulated through all the data, it will present a summary specifically and explicitly citing the sources that it came from.

Excellent. Here we go.

Great. So as we can see, when it's coming in here, we're starting to see the streamed answer basically underpinning that it's complex, and we expect the conflict to continue well into next year. This is not something that any of the experts on any of our teams expect is going to be a 2024 resolution. And you can see there, once it produced that short, actionable summary, it cited the 2 documents that it found it from. I think critically here, as you see, it didn't pull just from one product suite. It pulled from both Dragonfly, a more tactical forward-looking intelligence product, and our market-focused FrontierView product to really look at this conflict from all angles. Okay. Let's take a look at another question. What will the likely economic impact be of the Mexican presidential election? All right. So as you can see, it's coming back with the likely impacts of that and specifically noting both the factors on the pro-business side and also some of the more populous policies we might experience from this next president. And then again, it's figuring out where is the right answer. So in this case, it's pulling from FrontierView again, as I mentioned, that's our more macroeconomic intelligence-focused product. And this time, it's pulling in a strategic perspective from Oxford Analytica.

Excellent. These types of answers are already having a significant impact in delivering our beta users value in a more timely and actionable way. Customers have specifically told us that this is saving them time and ensuring they find information they might otherwise have missed.

The Copilot for Global Intelligence is currently available in beta to select FiscalNote customers or prospects. But I'm excited to announce that next week on July 2, it will be going to full release to all customers and prospects.

I'm really looking forward to the launch, Josh.

Thanks so much for all of your time and attention, and I'm excited to see what the next session holds.

Next, Josh Haecker, Head of Product for Global Intelligence; and Regent Armstrong, Director of Data-Led Advisory for Global Intelligence will go over Risk Connector.

Hi, again. I'm Josh Haecker, Head of Product, Global Intelligence. FiscalNote Risk Connector is designed to help organizations and governments manage their supply chain and vendor management risks. We launched Risk Connector about a year ago in response to customer demands for better, faster and more data-driven ways to map their supply chains and identify risks while they were still actionable. Think of managing a supply chain's risk like playing chess. If you're trying to play defensively, you want to anticipate what your opponent is planning and position pieces appropriately to thwart them. But for supply chain managers, thinking about risks the same way, the status quo is they can't see the whole board, and it only updates with the 5 move delay. That's basically an impossible position. Now imagine not only being able to anticipate your opponent's moves through the end of the match, but also getting data about every other player in the tournament. That's what Risk Connector does. With Risk Connector, we give visibility into an entire supply chain and everything happening within it so that critical business decisions can be executed with speed, confidence and foresight. In addition to lacking necessary data and technologies, other factors are increasing the vulnerabilities of supply chains. I'd like to invite my colleague, Regent Armstrong, our Director of Data-Led Advisory, to tell us more about those factors.

Thanks, Josh. We've never heard from our customers more about the critical challenges facing supply chains. With rapidly rising geopolitical tensions and expanding international relations, the environment is simply forcing their supply chains to shift often without their full knowledge. We're also hearing of the rising impact of seemingly small or local issues that ripple through supply chain and cause major disruptions or the impact from a company they didn't even realize they were connected to. One example that many of our customers cite back to us is around the collapse of Silicon Valley Bank. While many of our major enterprise customers did not rely on SVB, they were unaware that, in many cases, the technology vendors powering various applications they use did. So when SVB collapsed, the resulting impact on their core business technologies and ability to operate caught them blind. Ultimately, our customers faced the problem of supply chains moving so quickly and have data becoming more and more diffused that they've struggled to keep up. I'll hand it back to Josh now to talk about how we approach solving these problems for our customers.

Thanks, Regent. The solution was really in thinking about how to gather enough data to find new connections and to catch those connections quickly enough to allow customers to act on the information. FiscalNote Risk Connector scours the Internet, looking not just at news and media platforms, but also structured data repositories, such as SEC filings, court document repositories, regulatory databases, et cetera, in order to comprehensively map and alert on supply chain shifts. Risk Connector then produces an enterprise-ready data solution that alerts customers within a day of potential risks to their supply chain, whether these be new entrants or events happening to existing known suppliers. They need to know where all the pieces are on the chessboard and what moves they are making. Checkmate. For our largest enterprise customers, they consume this data wholesale and then integrate it directly into enterprise risk management or vendor management solutions that they've already purchased or developed in-house. However, for companies that are still developing risk management solutions of that scale and cannot integrate a raw data feed directly, we partner with Regent and the advisory team to deliver more actionable qualitative version of Risk Connector.

Our goal really is to leverage our team of experts and advisers to make the Risk Connector data as actionable and tailored as possible. In addition to providing a simplified map of their supply chains, our teams will also do bespoke analysis to identify risks and opportunities, run impact analyses, create action plans and regularly monitor their supply chains for changes. This dedicated advisory support lessens the load of our customers and allows them to act with confidence that they shift when they shift sourcing patterns, adjust how they support different markets or even just proactively begin a messaging campaign around a nascent PR risk. The most common way we deliver this support is through customizing and maintaining a dynamic dashboard. Josh, would you like to walk us through an example?

Absolutely. Thanks, Regent. So as you can see here, I'm looking at a dashboard for company A. I should say before I go through this and the next example we'll do that all of this is fictionalized data. You may see some real company names that we've pulled just from companies that everyone knows, but none of this data is about any company's real risk or any company's real supply chain. With that caveat out of the way, what we can see here is a dashboard that immediately bubbles to the top from millions of rows of data from the full Risk Connector data feed, the top 5 most impactful risks to this company. In this specific client engagement, when Regent and her team sat down with them, they also determined that the customer had a major fear of unanticipated PR challenges. That's why the other thing they wanted us to prioritize at the top was the top 5 risks based on sentiment around the vendors they're choosing to work with. And as I scroll down this dashboard, you'll see as I hover over it that it's dynamic and has additional information, we also see saliency around how relevant the suppliers in their supply chain are when they're having risks, and we also see the ability to concentrate those risks by theme. Again, boiling this ocean of millions of lines of scoured data up to the things that are actionable and matter. And what we can see here is that there's actually a pretty high concentration of cybercrimes, other cyber-related security issues and lawsuits throughout this customer supply chain. Finally, if I scroll down to the bottom, we can look at the risks by some individual companies. And again, these aren't actually any real companies, just some names you might be familiar with. Here, we can look at those various indicators on the level of impact of that company that they have on a supply chain, but also the confidence that another risk is imminent. And by using these data points, they can make actionable data-driven decisions.

While the dashboard is a more dynamic way to consume the information, often, our customers also seek our advice on how to act on the information and put those insights into practice. For our customers, we will produce a periodic report on key risks and what they mean specifically for their business. We tailor these reports quite closely to customer needs during stakeholder workshops at the start of the engagement. Let's take a look now at a report.

Thanks, Regent. So here we are again with our fictional company A. And as Regent said, the goal here really is to boil this down to normal language. So we want to take those billions of rows of data and actually get even simpler than a dashboard with various sentiment scores and impact scores and just talk about you're having major issues with cybercrime, sanctions, lawsuits and bankruptcy throughout your entire extended supply chain from the third tier to the fourth tier to the fifth tier. And as you can see in some of these lines, we're specifically calling out which company impacted which other company in your supply chain and that there's likely to be a significant vulnerability or a knock-on decision that could dramatically impact your ability to source the goods and things you need. This looks extremely simple, and that's actually its strength. And that's why we do so many long stakeholder workshops at the start, so we can figure out how to take all of that data and make it actionable for our customers right away. Risk Connector is a clear demonstration of how we think about melding human expertise and consultation with AI-supported data products to deliver fast, broad and actionable intelligence to our customers. Thanks so much for your time and attention, and I look forward to seeing what's in the next session.

Next is Andy Chakraborty, Head of Financial Products and Data Science, who will go over StressLens.

Hi. My name is Andy Chakraborty, and I'm going to tell you about how FiscalNote is helping customers solve the problem of people not always saying what they mean or meaning what they say with one of our newest offers, StressLens. Much of risk really can be thought about being about other people and the uncertainty that they give you a world. FiscalNote itself has been looking at lots of different kinds of unstructured data, natural language processing around news, information and as well, things that are incredibly anti-structured, things like SEC filings. But to really understand risk and people, you have to look at the people themselves. Much of AI is really about treating computers as though they're people, augmenting and supplementing human intelligence. However, we shouldn't go the other way. We shouldn't treat people as though they're machines. So the really only way to drill into what people have going on is to look at the individual themselves. When we look at the words, we get only 7% of the information. What's the rest of the information? Well, it's stress. When we talk about stress, we mean a lack of control or certainty that people have around the subject that they're talking about. You can argue that, that's the most important thing that you want to take a look at. How do people feel about what they're talking about? How much control do they have? How are other individuals reacting? How are the questions that they're receiving affecting them? And in fact, you do this in your everyday life, with your best friend, with your child, with your mom, if they ate all the cookies, if they got a new job, what is it that's going on? What are you doing? You're looking at a bunch of nonverbal cues, things in their voice, things in their facial expressions, and indeed, our product really just does that. It enables the software and the AI to scale that to people that you don't know. So let's take a look at how this works. This is a journalist with CNBC asking the Chairman of the Federal Reserve, Jerome Powell, a question about monetary policy. So as you see, we're tracking things like all of the landmarks on their face, the eyebrows, how they hold their eyes, their tone of voice, how much sweat they have, their pulse, their respiration, things like that. Why would you do that? Well, you really want to understand not only what words the person is saying, and you can see some of the words -- the words have popped up and you can really understand some of the topics that they're talking about. But you really want to get a sense for how do they feel about it. And as well, there might be signal in the questions that are being asked or how are people reacting in the audience or other people that have jurisdiction or relevant information. So really AI, what is it doing? Well, it's really helping you solve the problem of what is the information content of this message. It really, in some sense, helps you avoid boilerplate. But I would argue there's another kind of boilerplate, an emotional boilerplate. You really want to summarize or understand or drill down into what the person is not sure about because that's driving risk or change. So that -- where does that emotional boilerplate come from? It comes from the thing from removing the things that they're not worried about and looking at the things they're worried about. Here, we see 2 individuals, Elon Musk and Mark Zuckerberg, well-known executives, who you can argue are similar to each other in some demographic respects. However, nobody would argue they're that individually similar. So really, by teasing them apart and looking at each individually, you can truly understand where one is uncomfortable or comfortable and what's driving that. We've been looking at this and using this toolset since 2017. So let's take a look at how we do this. Here, we see an example from our data explorer tool. Let's switch to FedEx, the package delivery firm. And you can see line by line, each of the individual text items that you can figure out what topic and what was the stress of that particular time during the call. And then as well, you have all the information that you need to join it to your other data sets. And as well, you can see that it has every individual, if you wanted to explore person A versus person B. So you see their stock price or equity price in green. Let's drill down on the lockdown period, the COVID lockdown period. That was significant for FedEx. Their stock price was on a run because what did we all do? We stayed home and bought stuff. So it's interesting to take a look at that phenomenon. But before we do that, let's talk about the people. In this case, let's talk about Chairman, Fred Smith, the Founder of FedEx. So he famously bet his fuel costs at a turn of the table at craps in Vegas to keep the company afloat. He was a forward air controller during his military career, a scary job, even he can't hide from StressLens. Now let's take a look at the story. So their stock price again is in green, and then you can see their stress scores from various earnings calls at the bottom of the chart. So the dashed line is the average trailing stress. The solid black line is this particular call, and then every individual dot is that individual comments you can drill it out. So here, we can see that their stock price fell suddenly, "suddenly" in September of '22. But if you look 3 months earlier, you would have had warning of the particular business issues that drove it. So it was about e-commerce share and fuel prices and general boring operational efficiency issues. So this is not just about finding catastrophes. Now what's interesting is it sort of works the other way, too. So imagine being 60% sure of something. You're probably not going to tell your boss or your customer because you don't want to get in trouble for something that was a possible benefit, right? You don't want to turn something good into something bad. So around Christmas time of '22, they were, let's call it, 60% sure that they'd solved a lot of those issues. And then you see in 3 months' time after their first quarter earnings call, the stock price return. So it's quite a round trip from the high volatility round trip in their stock price. But what could have happened is with StressLens, you could unpack that, again, looking by topic, understanding what are the things they're not sure about, understanding when they feel like those issues have been fixed. So we've taken a look at drilling down in a business area. We had a customer ask us, "Hey, can I look at this in the geopolitical arena?" And we did so in concert with our friends at Dragonfly. So here, we have a concern index. And like every other index, it's just sort of a normalized view. In this case, it's the incidence rate. So talking about geopolitical or public policy problems and then [ it's, well, your] stress associated with it. This is 5 select sectors, pharma, regional banks, things like that. And you can see the shaded area around them talks about how much variability there is. Sometimes interesting to look at, again, what the opinion is. Is it consistent opinion? Is there widely varied opinion? Is it going up? Is it going down? So here, we can see a few different spikes, which are really interesting to drill down into. I'll just tell you one story in the interest of time. Here, we can see regional banks, 1.5 years, 2 years before all of the events of Silicon Valley Bank and other regional players. So you can see them starting to tick up once the issues around the COVID lockdown started to become material, things about debt service payments and operational issues that those banks were seeing as dramatic early warning indicators. Now it is interesting to note that we're seeing some of those self -- those same flags showing up this year, an interesting story to drill down into and a little bit of a mystery. Now let's look at really kind of the last story of these things coming together. I told you StressLens is about people. And what's critical about the human animal? It's a social animal. So because of that, we want to understand how things relate to each other, how company relates to its suppliers, its competitors, the rest of its sector, how one sector relates to another sector, what issues move the needle. So here, we see a statistical relationship among sectors, but what's critical isn't the statistic, it's the story. So if I'm happy when you're said about a particular topic, and we're connected, right, we have important business relationships with each other. One of 4 things is true: I have information that you don't have; I care about the issue more or less than you do; or our interests are perhaps not aligned in this issue; or something is changing. I would want to drill down into any of those issues to understand either in a product sense, in a geopolitical sense and a public policy sense, what's going on. So we've seen kind of how this works both in monetary policy, in earnings, in geopolitics and global security. In a little bit, Jason and Bill will tell you about how we do this looking at politics as well, where we've been applying this technology for quite a while. Thank you very much.

Next is Jason Dick, Editor in Chief for CQ Roll Call; and Bill Frischling, distinguished scientist and Vice President of Emerging Technologies, who will be discussing Roll Call, Factba.se and StressLens.

Hi. I'm Jason Dick, and I'm the Editor in Chief of CQ and Roll Call, and I'm going to discuss a little bit about how Roll Call and Factba.se have combined to complement award-winning journalism and deep historical data. Roll Call was founded around 1955. We're founded by a man named Sid Yudain, who saw a need to focus a little bit more on some of the politics and the people who make things run in Congress. We spend a lot of time covering the inner workings of the capital of legislation and the people who make that happen. And we thought it was a pretty natural extension to expand some of that, the coverage of the White House and some of the things that Factba.se does so well. Without further ado, I'd like to introduce my colleague, Bill Frischling, who's the Founder of Factba.se, and he's the distinguished scientist and VP of Emerging Technologies here at FiscalNote.

Cool. Thanks so much, Jason. I appreciate it. I'm really excited to tell everybody about what we have going on with bringing together Factba.se on Roll Call and what it means not just for the company, but for the elections, for political discourse and for society as a whole. And I don't say that lightly. We were founded about 7 years ago with the goal of becoming the canonical record of every single word spoken out of the White House. It was a stretch goal when we started and we've succeeded beyond our wildest dreams. Fast forward about 7 years, and we're now part of FiscalNote. And we are relied upon day in, day out by the entire White House press corps, by the White House, by most of Washington. If they want to confirm whether or not a President or a presidential candidate said something they know if they see it on Factba.se, they know for sure, it's absolutely true. It's a reputation we stand behind, and we're pretty excited about it. One of the things that makes Factba.se particularly unique is that we decided from the very beginning that not only would we avoid any commentary, we'd avoid any hint of commentary, and we left it entirely to AI analysis. This is going back a few years now. Fast forward now, and what we have is an extensive database that doesn't exist elsewhere, that allows us to analyze the tone, the content, what people are saying, how they say it, and lets people see not just the reliable record that Factba.se provides, but also see data and analysis that is impossible for anybody else to do because of the way we built it and the data that we have. That's what makes bringing this together with Roll Call so exciting is it's taking this data set and merging it together with a world-class journalism at Roll Call and allows us to do some fairly spectacular things. So if you don't mind, I'd like to take a quick moment and just show you a little bit about what we have going on with joining these 2 products together. So let's start with the basics. Most people find Roll Call Factba.se when they're googling. They'll find us usually at the very top. And usually, they're looking for something particular about Joe Biden and or Donald Trump. Of course, we have a high level of trust. We usually show up pretty high, which is kind of nice. And that's how they will usually have their first experience on the site. But we've added a lot of new features that really help you drill down into the data. So we have over 200 different data points for every single one of the transcripts that we have sitting in our database, the transcripts and the video. And what you can do is you can search for things like topic like immigration, but you could also drill down. You could sort it down by based on the topics, based on the entities, based on the sentiment score, based on audio scores, based on location, based on -- I can keep going on and on, just about any way you want to slice the data, you want to look for Republican Congressional districts versus Democratic. We have all of that, and you can -- just a click away, you can filter the information down. This is a really powerful tool because it just gets you a lot faster to exactly the data and the information that you're looking for. One of the more powerful things that you could drill into using this data is one of the facets is StressLens. And StressLens, I think as touched on earlier, really comes down to the who, what, when, where and why. It is the absolute crucial piece that really anchors everything that we're doing in the relaunch. So I'd like to take a minute here and talk a little bit more about StressLens. Everybody from about the age of 18, their speech patterns are locked. How you say your comfort level, your rate of speech, your level of stress, whether you're on script or off-script, whether you're on camera, you have a certain comfort level, and that just largely gets locked in. And now picture that pattern when you have literally thousands and thousands of hours of data on, say, Joe Biden and Donald Trump. And now you can look and see not just what they're saying, but what they're trying not to say. You can see what they say and you can also see what they actually mean. You can look at how, let's say, Joe Biden speaks about immigration or how Donald Trump speaks about immigration and compare it across decades of history and look for deviations and find patterns. We find patterns as well. You can dig into it also. This is an incredibly powerful tool that has applications not just obviously for the White House, but for all of FiscalNote, for all of our clients and for all of our readers. We're excited to be able to provide it for free through Roll Call, and we're also excited to continue to provide it to clients and continue to expand this. So you can see how this appears when you start drilling into our new transcripts page, where we put a lot of effort into making it a lot easier to be able to navigate around within the transcript itself, but also to be able to find all of this extended data. So right at the top, we have the StressLens score where you could drill in and see where the deviations are occurring and chart it out as a time series. You could also drill in, as always, in reading through, clicking on anything. You'll see this is an interview, a fairly recent interview that Donald Trump did with a podcast. You can jump right in. And if you were to click on any one of these, it jumps right in the video to the time code. If you were to click into one of these paragraphs, you can drill into more than 100 different individual pieces of data that we have on every single paragraph. If there's something that you're interested in learning about, you can find it. We also we also go through the process of making it easy to find these particular deviation. So any -- within any particular document within any particular transcript, we highlight automatically where the deviations are larger, so you can jump right into those pieces. And this is something that is just the start of what we're doing, and it's particularly powerful and important for the selection season to be able to find not just this information, but to be able to have this kind of targeted data in order to be able to drill in and find essentially what you're interested in drilling into knowing that you're getting it from a source that's trusted by pretty much everybody who covers the White House and the White House itself. So it's not just these things. We also have the schedule for the President, and we'll soon have the schedule for the [ track ] campaign as well in here. This is one of our more popular features. If you search for it right now, and feel free if you're out of compute to do a Google search and look for a bite and schedule, we show up above the White House. We're actually the top of all Google searches. That's the level of trust that exists within the product. We also maintained full canonical records of all social media, all press releases, all of this tied cross-referenced, tagged and filed and sorted 9 ways to Sunday throughout here. And this is just the beginning of what we have. As you can tell, I'm pretty excited about this. I've been working on it for 7 years. Hopefully, we'll be working on it another 7 more. And I'll stop here because I could keep going for quite a while. This is where a Roll Call working with Factba.se really enhances our ability to create unique quantifiable stories from a lot of all these different ever-evolving databases that you've been talking about, though. I mean, for instance, [indiscernible] is going to be a huge topic in this campaign. We are right around the 2-year anniversary of the Don's decision. And looking at the time frame for how Donald Trump and how Joe Biden have discussed this topic and in what settings, whether it's at a fundraiser or whether it's at a rally, that's just something being able to have that at the -- at our fingertips. It's just really valuable. It's something that it would just take an immense amount of time to compile that without this. Another example, not to talk too much about the Supreme Court, but the Supreme Court has issued 2 major gun decisions just in the last month. If we just wanted to block off the time from that first decision to right now say and see how Donald Trump and Joe Biden have talked about guns in that time since those 2 decisions came out. I mean, that's a very easy thing that we can do with fact base. And again, it's just -- the time saved for a reporter to look at these data sets is just something that we haven't had any kind of access to before. So it's really great. I'm really looking forward to it. We're excited.

All right. Thanks a lot, Bill. This is really exciting. And now we're going to go to Gerald and Fallon. Thank you.

Next is Gerald Yao, Chief Strategy Officer and Co-Founder, along with Fallon Farmer, Principal Data Scientist who will introduce Copilot for policy.

Hi, everyone. I hope you've been enjoying AI Product Day so far. I'm Gerald Yao, Chief Strategy Officer and Co-Founder of FiscalNote. I'm joined by Fallon Farmer, Principal Data Scientist and the Lead for Copilots. My co-founders and I started FiscalNote 11 years ago, as Vlad said earlier, the first product we built helped monitor U.S. state and federal legislating technology. We started as an AI company from the beginning, launching the first commercially available predictive analytics for legislation in 2013. While we may have grown to sort of a few thousand more customers since then, we've remained true to our roots, leveraging technology to help our customers get the information they need to make decisions quickly and confidently. FiscalNote Copilots are the latest step in fulfilling that mission.

Hi, everyone. I'm Fallon Farmer. Today, we're excited to introduce FiscalNote Copilot for Policy, our newest tool designed for people who work with legislation, especially government affairs professionals. We created this AI platform leveraging trusted data from FiscalNote to help users understand and act on legislation quickly. Over time, we'll continue to integrate these capabilities across FiscalNote's existing suite of policy products.

It's easier to show than to tell. So with that, let's jump into a demo. So, let's say, you're a government affairs manager at a sports gambling organization. You ask copilot a specific question on a specific bill. And it'll provide an answer based off of what you're looking for. So in this case, I already have a piece of legislation in mind. So what are the key points of Minnesota's house filing 5274? So copilot, as you can see, is starting to generate an answer. It's reviewing the request and it'll -- searching through fiscal database, looking at the actual bill itself and then reading it and providing key points on that piece of legislation. So it's like having an assistant that can read through basically any U.S. legislation and come up with analysis, summaries, action points, or anything else that you might be looking for in terms of analysis or finding a piece of legislation.

All right. So as you can see, Copilot for Policy has found the piece of legislation and it's streaming the key points of the piece of legislation. So with that, you can see that this is related to sports betting and fantasy contests. There would be new tax rates starting in fiscal year 2027 if this bill passed. You can see it's related also to horse racing. And if you're interested in seeing the direct source from the Minnesota legislature, you can also click directly on copilots to the site itself and see further details about the actions and the timeline. So going back to [indiscernible] Copilot for Policy, just skimming through the key points here, I am interested in seeing the potential impacts of public safety, taxes and commerce specifically. So with that, Copilot for Policy actually generated some follow up questions for me, and I just clicked on one of them. I hit enter and now it started to search through the Minnesota House Filing 5274 and start to analyze the potential impacts on public safety, taxes, commerce, state government, finance and policy. So as you can see, it's starting to respond one by one in bullet form on public safety, on taxes, just everything in terms of these areas and how this piece of legislation, if passed as is, would impact these sports betting and fantasy contests in Minnesota. So as I'm reviewing what it's streaming, it looks pretty good to me and I'd like to at the end edit this as a report and send it to my stakeholders, given the wide ranging impacts that it has. So when I click edit a report, you can see that there's text editing features already within copilot. In this case, it's the same output as you saw in the chat. It includes the source, and there's not much that I actually want to change here. So I'm just going to change the title of the note and also save it so that it represents the piece of legislation that I was looking at. So with that, Copilot for Policy is not only limited to a piece of legislation that you already know about, you could, for instance, search for sports gambling in Alabama. So my original query that I had last week was about can you tell me about recent sports gambling legislation in Alabama? And so Copilot for Policy looked at the most recent piece of sports gambling legislation. It summarized it for me and it also provided me the sources for all 4 pieces of legislation. And so, as you can see, Copilot for Policy is easy to use. The foundation of Copilot's knowledge is FiscalNote's legislative database, which is a trusted source that we've had and we've been building for 11 years so users can have confidence in its accuracy and timeliness.

Exactly. We're not just scraping information from the whole web. FiscalNote Copilot is purpose built for the policy world, drawing upon our own real-time data pipelines and document stores. Of course, as an AI system, it's not always going to be perfect. That's why we include links to authoritative sources as a reference. For comparison, if you were doing the same workflow with a generic AI tool, you would experience some key limitations. First, a generic AI tool's knowledge is based on a model that may be outdated or web data that may not be tailored to the policy domain. Second, it doesn't have access to FiscalNote's proprietary legislative datasets that power our copilot. And third, it's harder to trace responses from these AI tools back to authoritative sources for further verification and research. And with FiscalNote Copilot, we've addressed all of these shortcomings, providing a targeted tool for policy professionals.

Thanks, Fallon. I know we're almost out of time for this segment, which brings me to an exciting announcement. As of today, we are opening up Copilot for Policy to the public. If you're involved in government affairs or legislative analysis in any way, we encourage you to sign up for the free open beta program. Since FiscalNote's inception, our goal has been to build more valuable tools for our customers. Today, we offer the most comprehensive policy datasets and have over a decade of experience in building technology products for thousands of government affairs professionals. All of this has led to our newest tool, Copilot for Policy. We're excited to see how you use Copilot for Policy. As you try it out, please let us know what you think. Your feedback will directly shape our development. We're committed to working closely with you to ensure Copilot for Policy is the assistant you need. Thank you for joining us today, and we hope to hear from you soon.

Lastly, we have Q&A with FiscalNote's Vice President of Corporate Communications and Public Affairs, Nicholas Graham; Chief Strategy Officer and Co-Founder, Gerald Yao; President and Chief Operating Officer, Josh Resnik; Chief Technology officer and Chief Scientist, Vlad Eidelman; and Chairman, CEO and Co-Founder, Tim Hwang.

Hello, everyone, and thank you for joining FiscalNote's AI Product Day. My name is Nicholas Graham, Vice President of Corporate Communications and Public Affairs here at FiscalNote. We hope you've enjoyed and learned a lot from today's presentations about our products and our product strategy. Now, as the [ cap student ] to our AI product a program, we are now going to have a fireside chat with some of our leading executives and innovators here at FiscalNote. They're going to answer questions about what we've shown you here today, what it means for our customers and what it means for our company. So let me introduce the guests on this panel before we begin. First, we have Gerald Yao, who is our Co-Founder and Chief Strategy Officer. Next we have Josh Resnik, who is our President and Chief Operating Officer. We have Tim Hwang, who is, of course, our Chairman, CEO and Co-Founder. And last but not least, we have Dr. Vlad Eidelman, who is our Chief Scientist and our Chief Technology Officer. Welcome to all of you. Thank you for being here for this fireside chat. So let's go ahead and get started with some questions for each of our guests here today. And I'd like to thank our guests who registered for this and provided some of these questions. Tim, let's start with you. What is the value of having a product day for FiscalNote? Now, by showcasing our products and our vision for our products and our strategy around them, why is this happening right now?

Well, first of all, I think that here at FiscalNote, we spent a tremendous amount of time trying to invest in our products and innovate for our customers. And in particular, as we've sort of seen a substantial advancement in technology such as generative AI, we have seen a pretty substantial demand for basically integrating those technologies into what we're doing every single day. And so as we've seen earlier in the day, we spent a considerable amount of time and resources in making sure that we're meeting those demands overall. And so we're trying to essentially meet the demands of our customers with the advancements of technology and really make sure that the market, our customers, our shareholders and our employees overall really know what we're effectively kind of building out for the market here.

Excellent. Now, Gerald, we've heard from you and some of our other colleagues here today about our latest copilot innovations. Very interesting stuff, but specifically, some of the copilots that you showed were about global intelligence and also our newest Copilot for Policy. What's next for the copilot strategy for the company and what is your vision for these AI agents as we move forward into the next decade of growth and innovation for the company?

Definitely. So, as we saw earlier today, we started with 2 copilots so far, 1 for policy and 1 for global intelligence. And so in building those 2 copilots, we've created a foundational kind of infrastructure and technology layer for us to actually be able to more rapidly, one, build on both of those copilots for our existing customers, but also 2, to build new copilots over time, whether it's for other customers within FiscalNote, or for brand new markets with new copilots. And so in tandem with kind of building that fundamental infrastructure that's extensible for building new copilot features, new AI features, we're building a constellation of skills to help our customers as kind of Tim referenced with anything that they're really trying to add value to. So a lot of our customers, for instance, are looking to find accurate data at the right time, in the right place, and then they're taking that all the way from analysis to generating reports and communications for their stakeholders. And so across that end-to-end value chain, we're building the skills to be able to make that easier and better than traditional software is able to do today.

Tell me a little bit more. I'd like to know about the foundation that was built and it seems to me a very accelerated record time. You're using that as kind of a copilot engine, right to innovate even more, that was an incredible feat for the company. How did you do it?

Well, it's all about the team. So we've been doing AI for 11 years, but we've also been building software products for our company and for our customers for the past 11 years as well. So as I mentioned, the Copilot for Policy session of the day, we have a couple thousand customers and we listen to their feedback and we're very focused on what they care about. And we directly kind of input their feedback into making better products and services for them. So it really is about the customers and the great team we have at FiscalNote. And it all combines to make better products.

Exactly. That's right. Josh, over to you for a second. So we've just heard a lot from Gerald about what he's working on. And part -- a big part of that is obviously copilots and a tremendous amount of innovation across the company and lots more to come. But how specifically, in your view, do these products interact with these new products, interact with our existing products? And what does that mean for our customers?

Sure. So you'll see this interaction in a few different ways as we move forward. And as Gerald was saying, it always starts with the customer, right? So we think about what problems can we solve for our customers and how best can we do it? And we've been deploying AI and technology to do that for the past 11 years. So what you'll see more specifically is you'll start to see features from some of the copilots that we have in place today start to weave their way into some of our core existing products so that current customers who are leveraging those products can get the value of some of those advanced features along with other features that they use today, like workflow solutions and other types of information search and discovery tools to get benefits where they can. So we'll start to weave these features in where it makes sense. We'll keep products independent where it makes sense. And then we're going to also take some of what you've seen today, things like our AI agent StressLens and start to extend that into other arenas, like deeper into policy, for example, where you can really add a different type of dimension to understanding and predicting where things may go from a policy perspective or a geopolitical perspective by taking the tools that we've created and starting to apply them in different markets and to different segments. So we'll start to bring these to bear in our current products that we have today. And as Gerald was saying, as we start to launch additional niche products going forward as well.

And just as a follow on to that, this is something that we are doing in terms of existing products and innovating. We're doing this in tandem, seamlessly. At the same time, we are synchronizing the strategic process of doing this together, correct?

Yeah, that's right. I mean, this is a way of accelerating our development across our entire set of products. So we've been launching AI features. Well, again, we've had AI built into all our products, right, for the past decade. Right. But even over the past year, we've started to launch more and more AI features in our core products as well. This is just a way of continuing to accelerate that innovation because we can do something like develop this legislative copilot independently from the core product, but build it in such a way that we can take some of the core features and functionality from it and integrate it back on the core product. So we're essentially accelerating on all fronts by building out these copilots and agents.

And that's exactly what we wanted to ask of you. And you had a great answer for that. Vlad, we all know that over the past few months, really over the past year, there's been a lot of chatter in the media sphere and in the industry about the 3 core issues of concern regarding AI, trust, integrity and confidence. This is something that has been on the minds of policymakers as well as people who are innovating all the products like we are. And I guess the question is, how does that factor into our strategy of developing products, and why is that important to FiscalNote our customers? Could you touch on trust, integrity and confidence in AI?

Yeah, there's quite a bit there. So I'm going to try to divide it up into a few different facets because it is something we think about both internally as we apply AI internally to our own workforce and our own ways of doing things, as well as in building the products that we were talking about today. So I think overall, on both sides, the legislative and regulatory landscape, both in the U.S. and the world, is pretty evolving at this point. So the EU is pushing ahead with the EU AI Act, which has a very specific set of cases that it describes, and how to mitigate some of the risks on AI. And the U.S. is coming along. Some states have already implemented new laws around certain kinds of activities. And I think on the federal level, the NIST regulatory framework is probably gaining the most traction. But we're monitoring that because it's part of what we do and what our customers care about. But even as that all evolves internally, we've already actually implemented AI governance, policy and ethics policy, for only, in which we evaluate which tools we want people to start adopting and how to go about being able to access those tools, because we really believe at our core that enabling our workforce to try these tools to make them part of their own workflow in a way that's trusted and we're comfortable with that does not degrade the trust in our products is really important. So we want to evolve with those. So we've already started to do that. We've put trainings together in ways that we actually train our own internal workforce and provide them these tools. And so that's kind of one aspect. The second aspect is in our own product development. And so in the product development, we've talked about this in all the demos I think we've listed, but I think there's a couple things that really make our products stand out and help build that trust and confidence. One, it's the proprietary data sets that we're both generating and collecting and cleaning and standardizing and normalizing. So we become essentially the authoritative source. So even for public data that we're collecting, some people actually come to us first, as opposed to the primary sources, because they know that we've not only collected it from the original source, but we've made sure it's correct, we've cleaned it, and we kind of have a stamp of a FiscalNote approval that authenticated, right? Exactly. In a matter of speaking, it is authenticated by our internal automated systems, as well as by our internal editorial analyst, professional services, who often touch up and help actually pull out some of the data that actually might have been incorrect on the source or might need to be combined in certain ways or derived from the source. And so both in the internal processes and internal workforce, and in the ways that we pulled it together, these datasets form the primary basis for which then we build the AI systems we're talking about, which then that data has essentially then the knock on effect of providing a better product experience. And then we connect it together with several layers of human review, internal tools, observability. And so as the world evolves, as AI tools get better, we're constantly keeping track of what's going on there, but we want to make sure that at the end of the day, the customer doesn't really care about how we got them the product and the ultimate result. They care that they can trust it and they come to us for fast, reliable information. That's what we're going to do.

Exactly. Now, Josh, I want to come back to you for a second, because there's an element to what we were just talking about with Vlad that I think plays into the next question. So one of the things that I wanted to ask you is why does FiscalNote and you and FiscalNote feel it's so important to ensure that human analysis and intelligence remains an integral part of the AI products and AI strategy that we are working on? And this is a big competitive differentiator for us. Can you explain to us why that's the case?

Sure. I mean, it really comes down to context, relevancy, and trust in the information. As Vlad was just alluding to, customers don't really care exactly how this information gets them. What really matters is what impact does it have on them? For it to have the impact that they need, it needs to be something they can trust. They need to know the information is accurate. They need to know that it's relevant to them within the context in which they operate. There's so many nuances in, in all the situations that people want to understand and act upon today. And so we have teams of researchers, analysts and journalists, and who have that added expertise that can bring that added nuance and analysis for our clients. And it's all built on top of not just the data sets that we have aggregated and created, but it's also all the content that our teams have been generating, proprietary content that we've been generating for the past 50 years, in the case of our geopolitical analysis, through Oxford Analytica, through the past almost 80 years, on federal government analysis from CQ Roll Call, as an example. And so the answers that we produce, whether they're straight out of our AI copilots and agents, or whether it's presented in the context of augmented human analysis and custom analysis that we might do, people can understand that it's built on this whole foundation of knowledge that we have, this foundation of data and information that's proprietary to us and comes with the right context and nuance, so that our customers know it's going to be relevant for them, it's going to be accurate, it's going to be something that they can truly rely on for their business, for their government, whatever the case may be.

Excellent. Thank you for that. Now, this is kind of a Co-Founder question, so I'm going to ask this jointly of Tim and Gerald, if you don't mind. Obviously, both of you are very keenly aware of this. Everyone's talking about AI. Everyone says they're leading AI, companies are talking about this AI product and that AI product, FiscalNote has led the AI revolution since 2013, not 2023. And so the question is, why should FiscalNote be chosen over the rest of all these companies who are claiming to be AI companies and have the best of the AI that they have in the marketplace. Why FiscalNote?

Yes, so I'll go first and I'll pass it on to [indiscernible]. So, as you mentioned, Nicholas, we've been doing this since 2013, and we were actually founded as an AI company from the beginning. So we've been thinking about this for 11 years straight without a break. And Vlad was actually one of our first hires all the way back in 2013 as well. So we've invested in AI. We've kept our team around, and as Josh mentioned, we've been building and collecting more of the data capabilities, the human subject matter expertise. So while we do have very strong capabilities in the AI, I don't want to just focus on AI alone is going to solve everything. We have intentionally built up all these other capabilities beyond pure technology, and then it's really the combination of all of that that's going to provide that trust and that reliability and ultimately, like, the value that our customers are seeking. And if a brand new AI company started today had a ton of resources, they actually wouldn't be able to copy what we do because they're missing that proprietary information, they're missing that subject matter expertise, and AI is not good enough today to actually, like, bridge that gap. So, Tim, I'll pass it to you.

Well, I guess what I would say is that we are already the market leader in this space. And so we service half the Fortune 100, we service hundreds of government agencies, 100% of the United States Congress, the executive office of the President, everyone from the federal reserve to the FDA currently consumes FiscalNote's products today. So authoritatively, we are the market leader in this particular category. I think when it comes to the artificial intelligence space, what we're seeing is happening, sort of an interesting explosion of different companies, right? So certainly there's the sort of foundation model game that's being played out in multiple different contexts. And on top of those foundation models, an explosion of MLOps companies and AIOps companies that are tremendously making progress based purely off of the demand for generative AI products. But I think sitting on top of the foundation model companies and MLOps companies, we're seeing vertical winners effectively, that are coming to market, that are leveraging their existing customer bases, their existing data sets to be able to be successful. Right? So you will see a medical AI leader, you will see a travel AI leader, you will see a tax AI leader. I think when it comes to our category, particularly when it comes to policy, legislation, regulations, geopolitical analysis, there is no other company that can beat us based purely off of the data that we have, the talent that we have, the thousands of customers that already trust us today, and of course, leading, building off what Vlad was saying, all of the internal tooling, the infrastructure that we built to be a trusted partner to our customers already, even if we didn't have any generative AI features, our customers today already trust us to deliver for them high-quality legislative information, high-quality regulatory information. I mean, the reality is that what we're really doing by introducing more generative AI features is really building on top of that trust and really taking them to the next evolution of what technology is capable of delivering, which are a lot of these skills and capabilities that we're bringing to market this year and of course next a couple of quarters as well.

Excellent. Gerald, I'm going to come right back to you for 1 second. You're sitting closest to me, so I'm going to pick on you a little bit. Let's go back to copilot solutions for a second. How does our copilot strategy address our customers pain points, since we know so many of the people who come to us for our products and solutions, their roles are changing all the time because they're incorporating AI into what they do. So the question is, how does that help them do what they do better as they integrate AI into what they do?

Yes, I touched on it a bit earlier, but I'll go even from a higher level for like what do our customers care about? Why do they buy...

That's a great place to start.

[indiscernible] products. So fundamentally, there are 2 questions that we try to answer for our customers. One, what's going on around the issues that they care about? And then two, like what should I do about it? And so FiscalNote, from everything that has kind of been mentioned already in the demos earlier, helps answers those 2 questions fundamentally. So, for example, if you're trying to find out what's happening in your landscape for your issues, I went into sports gambling earlier in the Copilot for Policy demo. You're trying to understand what's happening probably, let's say you're a global company, you're trying to understand what's happening across the world for legislation and regulation around sports gambling. And so FiscalNote has been collecting that data, has been cleaning that data, and it's also verifying that data with not just machines and scraping, but with humans as well, to provide what is actually going on in the world and in the issue of sports gambling. And then you go into the kind of the generative AI features and what generative AI is really good at is pulling out insights and summarizations and key points and next action steps information and what FiscalNote, as we have information on all of our customers, we're going to be able to customize those action steps and those reports with our customers to be able to make it more valuable and faster and better than ever before. So really looking forward to the releases that are coming out, not just today, but also in the quarters to come to be able to deliver that customized value faster, better, cheaper, some combination of those 3, again, in a trustworthy and reliable way.

Sure. No, we're all looking forward to that, too. Josh, I'm coming over to you next. How do you see enterprise adoption of AI being a core part of our strategy moving forward? And where do our strengths and competitive differentiators lie in that regard for FiscalNote?

Well, as Tim pointed out, we work with a majority of the Fortune 100, so we have, and beyond that, we have thousands of customers. So we have a tremendous enterprise customer base already. The reason why they're with us today is because they know that they can trust the information that we provide, that we provide them with the tools to then act on that information. So to Gerald's point, it's not just about what's going on, but what can I do about it? And so they're already trusting us with this type of information. They're trusting their businesses with us. They're trusting -- if you think about the government organizations that Tim mentioned, right, executive office of the President and more. So the most important decisions being made around the world are being made based on the information that we provide and with the platforms that we provide. And so as we think about going forward, what this means when you start to introduce the newest generative AI capabilities in that context, again, from a customer perspective, it's really the same need, which is what's going on and what do I do about it? And it comes down to who do I trust to give me this information? If I'm the Chief Risk Officer at a Fortune 100 company and I need to report to my board on an issue in my supply chain or an issue of global regulation that may impact my company or whatever it might be, I'm sticking my neck out when I go give that report to my board. And so I really had better know that I can trust the information. And so we're already there with that, and we're taking that same expertise, that same knowledge base that we have, and all the same skills that we have and extending it further with generative AI, taking care to make sure that the information that we provide is accurate and also with the ability to bring augmented analysis through human insights that we can provide through our analyst teams and research teams that we can help bring to bear to solve speed specific problems for specific customers. So this is something where we see tremendous opportunity as we look to deepen the relationships that we have with enterprises today and extend deeper into the enterprise market.

Excellent. Dr. Eidelman, I'm coming to you now and read a submitted question, if I may. Increasingly, multimodal assets such as alternative audio like podcasts and social media, are supplementing text and transcripts as important to our customers. FiscalNote has made multimodal approach an integral part of our AI strategy. So how does this differentiate us, Vlad, and provide us with a competitive edge again and an added advantage in the market?

Sure. So if you look at the trends, I think just over the last 20 years, more information has come from non-primary sources, like news, social media, other places. And all of that feeds into the context of when I want to make a decision about something, I don't think about different data sources as silos, I think of it as pieces of a puzzle that I want someone, maybe sometimes it's me, sometimes it's a program or application I use to piece those pieces of puzzle together to give me insights and answers so that I can do something about it.

Because multimodal means multifaceted, right, to have that approach.

Right. Multimodal means different kinds of modalities, so text, audio, video, different ways information is presented. So you and I are speaking, there's an audio stream of this, there's a video stream, there's going to be a transcript. The information in those is going to be equivalent, but presented differently, and machines work with that information differently as well. And so there's a realness, a real timeness to information that's a video or an audio transcript or kind of a conference or a podcast that has not traditionally been represented as part of the political regulatory [indiscernible]. People traditionally focused on databases of the primary source material, of the laws, of bills, of other regulations, of geopolitical market analysis. And so that's great. That's obviously fundamental to everything we do. But the more of these discussions are happening in other places, the more that Senators appearing on a news channel or podcast, the President appearing on a podcast last week. We want to make sure that we're capturing all of this as part of the conversation, doing the kinds of analysis on it that we've done with all of the fundamental data sources before, so that we can provide that to our customers, so that they have a complete picture of the world and not just limited into these different silos. And that's part of the competitive advantage, I think, is having multiple places that we're putting data together that compounds the value of it, and that creates something a lot more descriptive and insightful about the world.

And just as a follow up question to that, how much more of the multimodal do you see sort of becoming important to our customers and an important part of the content that we integrate into our sources, our databases?

Yes, I mean, I listen to podcasts all the time. I listen to the radio coming in here. I listen to various videos online about stuff. I interact with voice assistants more and more, as opposed to searching in more traditional activities. So I think that our customers are just like you and I, right. They're trying to find the most cognitively interesting and kind of low barrier ways of getting information. Because everyone's busier, there's more information out there. So the more we can actually capture it and reduce it down to something that's meaningful, I think the better off everyone's going to be.

Tim, we talk a lot at FiscalNote about being a product led and product driven company. What exactly does that mean, in your view, and why is it important to our business and to our business objectives moving forward?

Well, I think that there's sort of 2 meanings behind that, right? So the first is functional, and then the second, I would say, is strategic. So from a functional perspective, companies have different modalities that they operate with, sort of, I would say a center of gravity, right? So certain companies are extremely sales driven, others are very marketing driven. Some companies are very well known to be design driven. I would say at FiscalNote, we are very product driven. And what that effectively means is that the expressions of our interest in entering a market or working with our customers on solving a specific problem are expressed through the products that we create for our customers. And so when we say we are launching x product, what we are actually also saying is that we are entering this particular market or trying to particularly service this customer. And so the expressions of our product driven strategy I think are very clear in terms of the investments that we're making. I think the second thing is, from a product-led growth perspective, is thinking about how we go to market as an organization. And so certainly we as an organization have our existing business development teams that work very closely with our existing customers, many of them large enterprise customers, government customers and the like. But as we introduce products like copilots and others, we're also now introducing ways in which users, individual users at companies, are capable of purchasing some of our products without [ owners ] procurement processes, or very large contracting processes. And I think that the introduction of that approach is going to be quite transformative for the company as a kind of new revenue generation opportunity for us as well. So something that we talked very extensively about as a company, and certainly the copilot kind of go to market motion, as we've seen it expressed at other companies like Microsoft or OpenAI or others, are being led primarily through the product led growth approach. And we are going to sort of see FiscalNote really come to market with a handful of products using kind of similar approaches as well.

Vlad, back to you. With things are changing so quickly right now, in part -- in no small part due to AI and technology in general. So how do you and your position, how do you ensure that our products continue to be cutting edge, that we continue to innovate and that the things that we are developing don't become obsolete so quickly? So what is the opportunity for us to tackle that challenge? How do you do it?

Yes, I think we think about this question a lot. I think about this question almost all the time. So there's a couple facets to this answer. I think first is this is a time for every company to have that question posed to themselves. So there's, I think, a wave of disruption that's happening. Some of it hasn't really reached its full [ zenith ] yet, I think a lot of it hasn't. But there's inklings of how it's going to affect almost every industry in certain ways. And so that's why all the things we just talked about...

Things are shaking out, starting to shake out.

Yes, exactly. And so I think as we're getting to the point of launching more and more of these products and having a real impact, it's really important to keep abreast of what's going on overall in the industry, both in our legal, regulatory, political industry, but also in the technology industry and the world of how both the ethical, the reputational, the regulatory and the technology kind of capabilities evolve, because all of those are going to affect how our products evolve and what can happen, what should happen. So we keep abreast of that. I think second point is that what we consider our core motes is what we're investing in to keep essentially ahead of what we see as rising capabilities across the board. So foundation models, frontier models [indiscernible] they will keep getting better. They will incorporate more multimodal communication, there will be more voice activation and more language and more images, all videos and coming together. And so we want to take advantage of those. That's not our core business. Our core business is not building the best foundation model. Our core business is applying the best technology that we can. And sometimes that can be us using open source tools, sometimes that can be using third-party tools and others to apply that to a lot of the, again, authenticated, trustworthy, reliable data that we've gathered and done work on or generated ourselves. And so that's kind of our first moat, is to make sure that our data roadmap of what we acquire and what we generate, how we derive metadata from it uniquely, is something that we keep investing in and applying these tools to accelerate that effect. And the second point of that is the actual way in which customers then use that, right. It's not about single transactional dialogue interaction, it's about positioning my analysis of the bill that Gerald was doing, for instance, in the copilot, in the context of what is my overall advocacy strategy, what is my regulatory strategy, how do I work with a compliance team, the other parts of the organization, how do I message that to my PR team? How do I incorporate that into our overall brand? And so all of those things are parts of an end-to-end solution or workflow or we are trying to position ourselves as having pieces that then come together. So there may be a single point solution or single transactional chat you could have about any one of those things. But we want to make sure that our products are actually solving for this end-to-end problem and growing and kind of increasing our ability to position that. When I say workflow, that could mean through a dialogue, right. The way that we interact with applications is also going to change. It doesn't have to be through another kind of different experience. But what I mean is, underneath that, it's actually the ability to do all of these different things in kind of a software suite that we want to empower. So those 2 things, data plus that kind of encompassing end-to-end solution. And then the last thing I think this is true for every company is keeping the culture of the team in innovation something that we consider, I think, paramount. It's having in the DNA, right. It's basically having every person be able to have an idea and be able to promote that idea and run with it and try to make that part of our culture so that I'm going to not have often the best idea. And oftentimes I won't even have enough of a contact with some of the evolving technology or the customers that I should be the one having that idea. But I want to make sure that everyone is able to then bring that quickly up and make sure that we can execute on it reasonably quickly so that we can push it out and test and validate. And that's just part of a non-product answer I think that as important as anything else I've said.

And I think, and, Josh, I'll have you just step in here really quickly. I think one of the things that we always talk about, too, that's an incredibly important viewpoint that we use to meet this challenge is customer input, customer feedback, listening to the voice of the customer. So I thought maybe you could touch on that for a minute.

Yes, absolutely. Understanding your customers is really paramount in all of this. I mean, I do want to echo what Vlad said about the type of innovative culture that we have, how the teams are empowered to act on things, how I do think we create a very good space for all the good ideas that individual team members across the company have. There's room for those to bubble up and for us to act on those. So I do think that's a strength of ours. But fundamentally, what's tied to that is understanding your customer. Again, we've talked about this a lot, even just over the course of this Q&A, but it really comes down to what problem are you solving for your customer and being able to think in a very far reaching kind of way about the expansive capabilities there are to solve those problems in innovative new ways. So it's really understanding your customer, which comes from that direct interaction, from listening to your customer, understanding their market. And so that's something that I think our teams really understand as well and helps to drive a lot of the innovation that we do. And as we start to bring new products to market again, we're very focused on not just developing them in a vacuum and bringing something to market from there, but it's how do we test this out and who do we talk to and how do we understand and get the market feedback that we need to take these good ideas and turn them into products that then can drive the right level of engagement, success, and revenue growth.

That's excellent. We're going to have to wrap things up. I do have a final question, and of course, it goes to Tim. It's been a great session, but if you could tell us, as we conclude this, how does all of this that we've talked about translate to value for the customer and ramp up our growth trajectory at the same time? And then I have to end with this question, what's the next big thing for FiscalNote?

Well, I think that just to summarize a lot of the conversation we had here, we've been investing in AI, particularly in this category, for over a decade. Now, when we say that, let's talk about exactly what that means. It means that for over 10 years, FiscalNote has built the tools and the systems to collect information around legislation, regulations, government policies from multiple different jurisdictions around the world. And then we built artificial intelligence capabilities that are cutting edge in terms of things like machine translation, summarization, et cetera, et cetera. And so that's even before this whole wave of generative AI had come about and before we've sort of seen this entire level of adoption, I think, in [ craze ] in the enterprise. Now what's changed in the last call it, 18, 24 months, I would say, are 2 things. The first thing is that we have seen, I would say, a stepwise change in terms of AI adoption as well as advancements. I think, secondly, we have seen enterprise and organizations much more willing to adopt more cutting-edge, generative AI tools within their organizations and their workflows. And so we've laid out a pretty extensive vision of what we want to get to over the course of the next several years. And it starts from the foundation of high-quality information we've been collecting, the tremendous amount of AI tooling that we built around that information, really building on top of the kind of authoritative trust that we built with thousands of customers that are trusting our definitive data and our content to make decisions every single day. And then what's next for the company is really thinking about how to build these additional copilots and skills for our customers to be able to leverage that information in new and innovative ways, right? So we showed off several of these new kind of copilots today in terms of being able to really think about our legislative data, our geopolitical information. We're now looking at other uses of our information in multiple different contexts. And then, of course, helping our customers to do a lot of other things with that information, right? So literally everything from memo drafting to summarization to whatever the case may be, and those constellation of skills, we expect to continue to be adopted by our customers and to really drive a new level of growth for the company over the course of the next 2, 5, 10 years and beyond. So I do fundamentally believe that this is a tremendous market opportunity, that we are the market leaders already in the space and we expect to expand our market leadership even beyond from where we are today. One last example that I'll give before kind of passing it back to you, Nicholas, is in the United States, there are 0.5 million elected officials, all the way from the President of the United States to the local city council member sitting in wherever you're watching this. Every single one of those elected officials needs to understand regulations, needs to draft legislation, needs to work with other local and state elected officials to try and pass policies. Imagine FiscalNote, who already has the market leadership and advantage in terms of legislative regulatory information, who already has the advantage in terms of adopting kind of core underlying infrastructure AI as well as generative AI and who already has the trust of the existing leaders in the marketplace, combine that with a new product led growth strategy focused on getting our products to market to a wider number of users, you can basically calculate the growth effectively of what we're capable of doing and of course, and more importantly, the impact that will have for our customers and the country really as well. And so that's what drives us every single day here at FiscalNote in being able to leverage the information that we have and support, frankly, the world's most important decision makers in making decisions around everything from what we're investing in, how we run our country, how we regulate ourselves as a society and be able to provide that information for all our decision makers overall.

That's a very inspirational and aspirational place to end. Appreciate that very much. That's all the time that we have for today's fireside chat, and I want to very warmly thank all of our 4 guests here, Gerald, Josh, Tim, Vlad, thank you very much for your answers and your insights on the those questions. To all of our guests who joined us, thank you very much for your interest in FiscalNote and for your time. We certainly hope you've enjoyed FiscalNote's 2024 AI Product Day. One final note, we invite you to provide us with your feedback, submit any questions that you have about today's program, but also share a video recording of what you've seen today with your colleagues. You can do all of that by visiting the following, www.fiscalnote.com/ai-day. Thank you again for watching and best wishes to all of you.