Forrester Research, Inc. (FORR) Earnings Call Transcript & Summary

It is 2024 live webinar. So 2024 is not a year to be sitting on the sideline. Generative AI is poised to change everything and you're right in the thick of it. Used 2024 to be big, to go big, to be bold and be willing to experiment. Now over the course of this webinar, Forrester analysts will highlight the year's biggest opportunities as well as how you can drive progress and exceed your business goals plus Forrester CEO, George Colony, will share his thoughts about the future of generative AI and then will also be followed by a panel discussion with the analysts on this webinar. So a lot of ground to cover. So without further ado, let's dive into 2024.

Hi, everyone. My name is Audrey Chee-Read, I'm a principal analyst here at Forrester. I research all things consumer behavior, consumer trends, consumer insights, and I'm excited to share our first prediction with you. So at Forrester, we have been on the forefront of the role, AI and generative AI is playing in our businesses, in our operations, in our technology. But the altitude that we often fly at with businesses may not be the same altitude the consumer's at. So generative AI, specifically when ChatGPT splashed into our lives over a year ago, saw a lot of skepticism. People were worried about how it was going to affect their jobs. They were worried about what it meant for curiosity and thought. Artists and creators were worried about what it meant for their work. And this leads us to our first prediction. That even amidst skepticism, 60% of skeptics will use and love generative AI, whether they know it or not. And as is the case, oftentimes with poorly understood technology that promises to change the world, people tend to worry about the ethics and human impact of things. At our Forrester, we have a really great research resource, which is an online community of members of consumers that we can pull and really dig into the why behind questions we have across different topics. We talked to them a little bit about generative AI and the different types of platforms and tools they might use and how they might describe it. And in general, what we see is that people generally question the morality behind it. The drawback is that there's a huge potential for fraud. We heard things like 1984 coming to life. There's no need for it. They describe them as tools that help people be lazy and not have to think for themselves. But at the same time, people also saw a utility. It was designed to help your task more efficiently. It's a must-have to make your life easier. It's like having a librarian, a secretary and assistant with creative minds rolled into one. It creates new words, sentences and other content that bodes well and is easy to understand. And underlying a lot of this is the idea of trust, and trust will remain an issue, especially this year when there are big events like the U.S. presidential election or the Paris Summer Olympics that are going to open up opportunities for bad actors. And what we find in our data is this, consumers see that 45% of them believe it will pose a serious part of society. And we've been tracking generative AI and consumer sentiment monthly across different countries. And this number has remained steady in the past 6 months. It's hovering between 45% to 50% in the U.S. 75% of consumers think that companies should disclose when they are using generative AI when interacting with me; and 31% with trust information provided by Generative AI. And I want to pause here because this 31% is really important to note because trust issues isn't just up the morality or ethics of what generate AI does but how good the output of what the tool produces. So for example, here, it's not only consumers who are skeptical about the technology, but top culture actors and writers are still cocooning it. If anyone here is listening is a Saturday Night Live fan. This is a sketch comedy show in the U.S. There have been quite a few sketches Ghost Rider Strike that make fun of the AI tools. Superimposing basically bad outputs of AI onto their actors and musicians. And this notion, this part is going to be very important going into 2024. As media pokes fun of AI capabilities and focusing on giving it a negative perception. And for first-time users, they are shying away from more frequent usage because they don't think the generative AI tool can do what they wanted to do. Whether I didn't really quite answer the question they wanted to answer or produced a poor image like this one you're seeing on your screen. And it signals one thing for companies and for brands, for consumers who are less familiar with the tech. That they don't have too many opportunities to really get it right. Nonetheless, consumers are still using it. For generative AI, it is no longer so much about building awareness. In this slide, we have tracked it as I mentioned earlier, across various countries and over time. And here, you will see the differences between the U.S. and the U.K. and the differences between June and December of 2023. We've seen awareness stabilize over the last few months, but adoption is still slowly increasing across different countries. And consumers are still learning how to use it right now. Currently, when we break down the task they use it for, they find generative AI to be more useful in task-oriented things like finding an answer or drafting and creating something. There's still a little caution and hesitation and skepticism and more complicated proactive tests, like planning a trip or giving advice or recommendation. And it's important to note because in this new year, while we see skepticism and issues with trust, we're also going to see people use it and embrace it unknowingly. That's because platforms and tech companies right now are really racing to embed generative AI capabilities into their platforms. It's not just the big tech companies like Samsung, Google, Meta, Spotify. It's even nontech driven companies, retailers like Walmart are all adding generative AI capabilities into the apps and products every day. So what's going to happen this year though is how visible and invisible some of these generative AI products will be, how branded it will be. It may be examples like Samsung powered by AI phone that just launched yesterday. Essentially, there will be ChatGPT functions that you can use offline on your phone or Google Search has added on Bard and ChatGPT functions or Spotify is using Open AI's technology to directly translate podcast episodes from English to other languages, all matching the original podcaster's voice and tone so that it can feel more authentic than traditional dubbing. And lastly, kind of on the right-hand side of the spectrum is Meta's WhatsApp, which is a messenger using generative AI to create stickers that you can use in your conversations. So for consumers, what does it mean? Generative AI has put AI on the map for consumers. It's been the consumer-facing AI tool and it's really helping them understand how to use AI. It's the on-site of new tools has democratized self-expression for many consumers created more accessibility into complex subjects and tasks. But what we're going to find is that platforms and tools will drive consumer behavior with generative AI, but for getting to build trust along the way is going to cost brands. And this is going to be a very important theme in this ongoing year. And lastly, when we talk about trust, it's not just about transparency. There are many different ways to think about building trust. It's human oversight. It's making sure you're on top of the evolving regulations. It's having a clear alignment with societal value. Its a partnership with diverse voices because bias is a huge challenge and an issue for consumers as outcome. And with that, I leave you with our first prediction here. Thank you everybody. [Presentation]

Good morning, good afternoon and good evening depending where you are in the world. My name is Martin Gill. I'm Research Director and VP of Forrester, covering customer experience and I want to build on what Audrey just told you about consumer adoption, and I want to share not one but two predictions for CX leaders. And these are linked but not fundamentally dependent on each other. So first of all, half of large global firms will experiment customer-facing gen AI. And -- but 1/3 of all brands will launch experiences that are biased, inaccessible or harmful. Substantial problems here. So what does this mean? Well, companies are going to launch. We believe companies will launch customer-facing gen AI to provide convenient, engaging digital experiences. No one wants to create harm or cause bias. But this is going to come back and bite teams that haven't fully vetted these experiences for the potential impact of the potential harm they can impact. So how do you mitigate for this? How do you plan for this? Well, ask yourself a question, just because you can doesn't mean that you should. Audrey shared some of this data already, but some 45% of leaders say their firm will apply generative AI to enhancing customer-facing experiences in the next 12 months. So you can kind of see where our prediction comes from in that kind of space. But as Audrey also alluded to, 69% of U.S. consumers and I think it's 71%, 72% of U.K. consumers believe that any public content that's created by AI should be labeled as such. Customers crave transparency, they create information. They want to trust it potentially, but they want to know what they're interacting with. They definitely don't want to be fooled. Now when it comes to CX and probably digital leaders and product leaders, AI is not new, generative AI might be, the era itself is not fundamentally new. The wider [ remit ] and bubble of artificial intelligence has been around for a while. So think about applications like automated mortgage approvals for a mortgage journey or credit checks, that kind of thing, that kind of capability has been around. Within that, we get to machine learning. So think about predictive journey optimization or modeling of customer segments to give next best experience or the next best action. That's not new. Think about deep learning, self-driving cars that can park itself, much deeper applications of systems that can learn. Again, not new, maybe slightly emerging, but it's not fundamentally new. What we've seen change in the last 12-ish 13 months is the emergence of generative AI, which begins to let us to have conversations. So I imagine, for instance, a chatbot that I can talk to that I can interrogate my insurance claim, talk to data, find out where the -- where I am in the process and have it talk back to me. We're creating conversational and immersive experiences in a way that we haven't been able to in the past. So generative AI moves us from a CX perspective, from orchestration, where we're manually having to put things together to conversation when we talk about journeys, orchestrated journeys to conversational journeys. And it will manifest in 3 distinct ways. The first is when you think about the things that we're building. Websites, apps, we're worried about clicks. We're worried about interactions, physical interactions, maybe physical locations full of workflows. That will change to customers having natural language conversations, possibly not in 2024, but this is the direction of travel. Customers talking to brands, hey, book me a trip to Paris. As teams, we have to predict, simulate and orchestrate with data, with models, it takes a lot of skills. We have to manually connect data together. We have to know where to look for things. That will move to a world where we can talk to our data. We can look at journeys like a language. We can ask interactive questions. Imagine being able to ask a question of your platform, where do female consumers aged 35 to 40 drop out the purchase journey. What that is, is that changes the nature of skills. We move from a world where we have data science, modeling, coding, being the language of how we interact with data, leading to know where the data is, to a world of democratized data access, where anybody in your organization can ask questions and get insights and we can scale the experience design. We can build more things more rapidly. So this unlocks an amazing amount of potential. Potential for companies for you in terms of how fast you can go, what you can create. Potential for vendors who can go to market in new ways, building tools, with new customers and potential for customers, both what customers get but also how we can understand customers the level of empathy, the level of insight we can get to about customers. But we mentioned about risk, much of that thing is going wrong. There are risks emerging that we're not even set to manage yet. We haven't even foreseen yet. This is a wider risk management framework. So I'm not saying in 2024, these things will come true. But why the green bar, the furthest extreme, looks at global systemic risk in global warming things like an environmental damage. We're not -- we may be worried about that, but we're not worried about those in this context. Zoning in a little bit to more societal risks, Audrey mentioned about election, some 4 billion people across the globe will go to election this year. So imagine the power of propaganda is information, erosion of social norms as we all face the right to vote. But again, as organizations, mostly we're not responsible for that. We may have some government clients here. But broadly, the things that we're talking about building here are in the yellow box. They're more at the individual and the company level. AI poses a genuine risk of driving inclusion, discrimination, producing a lack of autonomy like an agency, privacy abuses, bias. Those are the things we as we design systems, processes, experiences we need to concentrate on. So what I needed to do, what I need to think about is look for 4 key risks as you develop AI-augmented experiences and ask yourself 4 key questions. The first is around bias. Does your model produce unbiased decisions and content. There are many examples of bias, you can go look at. Ask one of the generators to generate a picture of a nurse. It's always a female. Executive, it's always a male. That's a very basic level of bias, but bias can go much more deep. So ask, is your model producing unbias content? Next, is it transparent? Can you explain how the model made a decision that impacts the customer? Explaining how it produces 2 pieces viral and whether they're different is one thing. But explaining how you made a mortgage decision or whether you gave credit or to credit away from a customer, that's important. You have to be able to explain that when customers ask. And then reliability or unreliability, how predictable and repeatable is your model if you give it the same information, does it make the same credit decision every single time. You have to understand how it's making those decisions and be confident it's making the same decisions. And then finally, privacy, and this is a Wild West topic, particularly. Is your model generating -- you're using or generating personally identifiable information. This is difficult because a lot of these large language models have been trained on open data on the Internet and have ingested personal information, pretty much by accident. So you're potentially bringing in a new class of risk to your organization that you didn't even have in the past and you didn't develop in the first place. So it's difficult. Now there are some legislations emerging in some places, particularly in Europe. So we have the AI act, which is coming. That's going to govern how AI scales across Europe, that we have existing frameworks like liability framework, like safety legislation and then emerging technical standards and specifications. You may have European clients, you may not. If you do, you're going to have to comply with these rules. If you don't, I would recommend look at legislation like GDPR, which set the global standard for privacy legislation. Other countries followed California or CCA, for instance, began to follow. So if you're worried about our governance, if you're looking for models about how to think about it, look to the emerging EU legislation. But what can you do internally? Well, our recommendation right now in search checkpoints into your design process. You've probably got a design process. It looks something like this, double diamond, triple diamond design thinking, start with an impact assessment. Will your generative AI project impact customers positively or negatively assess that. And then put checkpoints in as you define the problem, as you define the solution, as you define the implementation. Insert user research into your decision-making, make sure decision maker is evidence-based, apply ethical research processes, create inclusive personas, use AI to help generate [indiscernible]. Monitor solutions as you roll them out. And as you implement them, absolutely test them before you inform them, but continue to test them, continue to cycle them and iterate based on what you find. Things are going to change. So finally, if you look at design thinking at all, you'll be very familiar with the concept of desirability, viability and feasibility. When we design solutions, when we build experiences, we're looking for something that hits the sweet spot in the middle. We recommend adding impact, what's the potential impact of your AI solution and keep humans in the loop to monitor that in the short term. So good luck in 2024. That's what I have for you, thank you. [Presentation]

Hello, everyone. I'm Laura Ramos, I'm a Vice President and Principal Analyst here at Forrester, and I write for a cross service team focused on customer engagement in the post sale. It's been my honor this year to lead more than 30 analysts in writing our 2024 predictions for B2B marketing, sales and product. Now like Martin, I've got a couple of predictions for you here. And we're going to focus on these two that are specifically around generative AI. They focus on product marketing and content marketing, respectively. So the first is that generative AI will surface insights that dictate 1 in 5 new B2B product launches. The second is that thinly customized generative AI content typically produced by marketers will degrade purchase experiences for 70% of buyers. On the next few slides, I'm going to share with you our thinking on why we chose these predictions and what you should do to anticipate the potential thrills and spills for each one of them. So overall, we expect 2024 to present B2B customer-facing teams with a thrill ride full of highs and lows. And AI in general and generative AI, in particular, is going to drive many of these twists and turns. In July of 2023, we did an AI pulse survey, and that data shows that 4 out of 5 AI decision makers believe it will have a high impact on their business over the next 2 years. Handling this degree of change will require marketers to become experts in this technology before fully understanding all the risks that it might expose to customers and buyers. And risks are likely because B2B firms are moving forward rapidly. From our survey data, we see that about 2/3 label themselves is experimenting or expanding their AI use already today. And an additional amount planned to significantly increase their investment there as well. But I think most importantly is that almost 80% say that generative AI will significantly impact, which products they're going to be bringing to the market over the next 2 years. Now when 24% of the generative AI decision-makers that we surveyed to see testing and simulation as an important use case. We see product teams poised to become an early beneficiary and this is why we're predicting that 1 in 5 products launched will be influenced by AI. Product managers and portfolio marketers must also intimately understand their customers end markets. And insights can be coming in from many different internal and external sources. So generative AI tools can help product managers rapidly sift through an even wider variety of customer-based information to find things like novel use cases that may represent a new market opportunity. By identifying these not so obvious capabilities, we believe that 20% of products launched will contain some form of AI-influenced features. So what do you need to do about this? Well, you're going to need to think more about continuous product discovery. And this is a practice that Forrester says is about identifying and refining customer value drivers through regular and interactive research activities. Generative AI is going to streamline continuous product discovery and make it more accessible to more teams. It's going to help them understand a broader range of customer needs, brainstorm new product concepts, accelerate product prototype evaluation and identify future needs sooner. So with all of that goodness, where should you start? What should you start doing now to take advantage of this? Well, we think you should use generate AI to identify product trends from more unstructured sources of information, such as competitive product documentation, industry white papers, wind loss reports, that sort of thing. Use it to process natural language in transcripts -- transcriptions that you might have from conducting customer research and then summarize the findings. You can also use it to streamline the production and review of things like release notes, documenting feature functionality and product announcements, things that typically have like a regular format to them. And then finally, use it in product relevant product documentation to make that documentation richer through things like AI-generated images, animated directions and even video presentations. Now on the flip side of the coin, we need to turn to our second prediction, which shows some potential stomach churning blows in to rapid generative AI adoption. Here's where we need to reconnect with our buyers. And in this data from Forrester's content preferences study, it shows that most buyers are really not happy with the content they see. If you summarize this, they want vendors to show that they understand their business needs and to show real value comes from adopting their solution. Now this has implications in terms of why we believe a lack of personalization is really going to degrade the experience. Most marketers 75% -- our data shows, 77%, believe that buyers do expect personalized experiences. But to deliver that, what we're seeing is that marketers are only really focusing on generative AI as a productivity tool. Now without that deep understanding of the customer working into the process, we're going to see that these early experiences focus on things like simple vertical veneers, poorly personalized offers and just ending up giving buyers more of what they don't want. And the result is going to be that the trend numbers we saw on the previous slide are just going to continue or maybe even go higher. I like using Lisa Duncan's quote here. She's the new Senior Director of Revenue Enablement at Meltwater. It's a media monitoring and analytics firm. And I like this quote because what she's saying strikes the right balance between productivity and precision. So when it comes to content that customers will see, marketers must ensure that not only is the content relevant to those specific customers. But it also delivers value in a way that stands out from the competition. So if you are a B2B marketer and want to know how to ensure that any content generated by AI hits the mark, then make 2024 your year for investing in fresh persona research. Start by asking your customer-facing teams, who are our best customers and how did they get that way? Use that research to identify the key moments that really change how customers benefit from using your products and services and the value that they realize in return. And like portfolio marketers, you leverage that AI tool to help with the research that you need to understand those personas and then guide the personalization that happens afterwards. So to summarize things just quickly. Whether these predictions come true exactly as stated or not. They show how important it is for marketers to invest in customer research, whether it's to make content production more efficient or to research new product ideas. Refresh personas will also help marketers ensure that they personalize content effectively for their intended buyers. I'm not just going to make the buying experience better for everyone. With that, I wish all B2B marketers everywhere, the best of luck in riding the roller coaster year ahead of us in 2024. [Presentation]

Hi, everyone. I'm so excited to be here with you today talking about our 2024 predictions. And today, I am going to be talking about 2 different predictions. One around AI, you've been hearing a lot about AI so far, but we're going to go a bit more specific on some of the technical predictions. And then we'll also be talking a bit about robotics and automation. My name is Rowan Curran. I'm a senior analyst in our Business Insights team, and I cover a variety of topics in AI and machine learning. And I would say, helm Forrester's coverage on generative AI from a technical perspective. So let's get into these predictions. So 2023 was obviously a very intense and kind of crazy year around AI overall and specifically around generative AI. And one of the big things that we saw was the introduction of a lot of proprietary models into the marketplace. But we are seeing an increased adoption of open-source models, and so we expect 85% of enterprises in 2024 to expand their use of AI and generative AI with open-source models. And this is being driven by a couple of different factors. So when we look at this, it's really can be addressed in 3 different -- looks at it 3 different ways. So the first one is that you don't necessarily need one of the biggest and baddest and coolest models to solve every single one of your enterprise problems. While the GPT 4s and the PaLM 2 and the -- in the -- sorry, in the titans of the world are very, very powerful and very useful, not every interaction requires that quality of interaction of summarization and their test generation. And so as folks have moved from their kind of initial prototype deployments from their first generation of applications, they're realizing that the power that they're infusing in these things really overestimates what is required for the function. And so that is one of the first reasons that they're moving to some of these open-source models, but they're also looking at this as an approach to cost control. So in addition to these models being overpowered for some use cases, they can also drive up the cost of implementing generative AI applications. Now this goes doubly for folks who are building applications that are being provided to some kind of customer that is external to your computers and some kind of end user. Where there going to be a high variability in the number of queries and in the volume of tokens that goes back and forth that can really drive up the cost of using these models, so we're seeing a lot of folks invest more to open-source to keep that a bit more under control. And the third reason here is really around where we can actually run these models. So if you're using a proprietary model from one of the cloud vendors, you're going to have to be running an inferencing against the cloud and sending some data back and forth over the wire. Now as acceptable for a lot of use cases for a lot of companies, but many of us have requirements around data security and privacy and things of that nature that even if we are working with a secure and trusted vendor, we can't send data back and forth to the cloud. And so we're seeing more folks bring in open-source models in order to be able to better control the environment around the model itself. And this also can potentially give you more control over the data and the understanding of how the model is functioning as well. And where is it -- this is actually leading us into is not an open-source dominated in a purely open-source world is actually leading us into a hybrid world. So we're seeing more and more open-source models appear on the market every day. And this is making it a very wonderful cornucopia to ingest, but it can be a bit overwhelming and you can get overstuffed. So there's a lot of challenges in this open-source world around choosing the right model. And so for many circumstances, it is still better to use a proprietary model for some of those higher end use cases. But then we're also seeing the model providers who offer these proprietary models also begin to offer more curated gardens and other environments to build with open-source tools as well. And so we've started to see in a large number of applications that people are starting to roll out is really a hybridized environment where you're combining open-source models with proprietary models in order to get the best effect for the lowest cost in the most efficient way. And just to give you a very quick idea of how this could look, I'm just going to show you an example of a retrievable augmented generation architecture, which can highly a hybrid approach here. So in a retrieve augmented generation approach, you have a question which goes into a system and has an embedding created. Oftentimes, folks will use an open-source model like BERT here to create that embedding because it is a relatively well prescribed test. There's a number of different models around it and it isn't necessarily the most challenging task for language model. Then in retrieval augmented generation, we send it back to the vector database, we get our data back. We match the scores. We don't have to worry too much about that. But then the next step is actually generating the answer. And so this is what we really care about, right? This is when it's giving that content back to the end user and it needs to be ingestible, understandable, we need to be able to make a decision of based upon that information. So that is where we're seeing more proprietary models introduced but we are also seeing folks use open-source model as you can see here. So we have both Llama 2 and GPT-4. And so what this amounts to is basically a more complex landscape of generative AI applications where you need to think about how you're actually going to be stitching together these different models to produce the best outcome at the lowest cost to really change the impact of your generative AI applications. So with that wonderful hybrid open world ahead of us, I'm actually going to pivot us to a little bit more of a clarifying prediction here. And that is around the automation and robotics space. And that's -- we expect to see complex automation projects crash at every hyperscaler. And this is really going to force enterprises to adapt to this reality of complex automation projects and how overwhelming they can be. So automation has been around in the enterprise for quite some time. But with hyperscalers, we're seeing more and more enterprises rush to engage with them to build automation, RPA projects across their enterprise at scale that really involve a lot of different complex pieces. And what we're seeing amongst a lot of these folks in this really reckless drive to automate more capabilities that it's leading to more and more failures of these projects, and that can lead to revenue loss, that can lead to brand issues. And so we expect to see more and more enterprises actually shifting to a more formalized recognition of how risk can influence their company and leads to negative outcomes. And this is being further exacerbated by the excitement around generative AI actually. So generative AI is great for enterprises overall. We produce a lot of really positive and growth-oriented effects. But if we look at this data from one of the algorithm incident databases that publishes every year. There's a huge spike in generative AI incidents in 2023. That's kind of to be expected. There was increased adoption. But that was at the same time, there was a reduction in a number of other incident categories. So I think it's important when we're moving into 2024 as builders, as marketers is all kind of enterprise, employees are responsible for quality AI projects that we understand that with complexity comes responsibility. And you need to be responsible for actually building in and understanding and a management of risk into your projects. And so we have 10 kind of core pitfalls that we suggest people to avoid in their automation projects. They fall around 3 main categories. Process, employee experience and customer experience. And really, that will help you get started to kind of delve into the solutions and to the approaches for managing the complexity in broad-based automation projects, especially when you're dealing with cross-enterprise connections in multiple different applications and they have different owners and they have many different pieces that are interconnected, interlinked. And so with that, I'll say thank you very much. I appreciate your time today and good luck in 2024 with all of your AI, Gen AI and [indiscernible]. [Presentation]

I'm George Colony, the CEO of Forrester. I'm going to talk a little bit about how CEOs should be thinking about generative AI for 2024. And why should this be a priority? This is the biggest tech change of my lifetime. And I've lived through a lot of tech change in the past. And I will say that in the past, Forrester's typically advise our clients to wait a little bit when the big tech change happens, but that's not the case with generative AI. We're telling our clients to move now. You can't wait a week. You can't wait a month, you definitely cannot wait a year. This is a revolution in the way we create, move and use information. This is a very, very big change, much bigger, I think, than the Internet itself. Now what is generative AI, I'm going to give you my very simple kind of CEO 50,000-foot view or definition of generative AI. Generative AI is a technology, which lets people converse and that's the most important word, converse with big piles of data and then helps them create something new and original from that big pile of data. Now what is the impact of generative AI going to be? Number one, we're going to move away from the web to a conversation with information, a conversation with big piles of data. So in the future, when you go to your bank, you'll go to a big prompt and you'll begin to converge with your bank. How much is in my bank account? How much do I owe in my credit card? How much is in my savings account? Again, we're moving away from instead of searching conversations. Two, moving away from search. This means we're no longer going to websites, searching, looking at all this stuff, trying to figure it all out for ourselves. We're going to again move away from search toward conversation. And then finally, number three, trust will be very critical as we move into generative AI. We cannot put out into the world, what Forrester calls coherent nonsense from generative AI information, which is really not very valuable for our customers. We have to make sure that, that information is curated. It's watched, it's built by human beings. I call this the -- this is not a robot moment where everyone gets to replaced by robots. It's what I call the Iron Man moment, where we put -- we all put on a suit of generative AI, and we're able to be more efficient and faster for our customers. But remember, the customer wants to know the behind the mask is a human being. So human touch will be critical in building trust for our customers as we move forward. What should you do in the next year? Number one, your company should have -- by the end of 2024, your company should have 10 ongoing generative AI projects. By the way, these don't have to be multimillion dollar projects that could be small. We should have 10 projects a year from now, one. Two, one of those projects should look as if it could be operational in 2025. So looking at the operational generative AI application either in 2024 or late or early 2025. So 10 projects, 1 operational. Number three, the IQ of your company around generative AI should rise for everyone. So this is not something. That's actually a really cool thing about generative AI. It's going to touch all of us. This is not for small segments of your company, just the techies or the marketers, this is for everyone. So number three, the IQ of the company should be rising around generative AI. And then finally, number four, the IQ of your executive team, your CEO and your Board of Directors in generative AI should also be increasing. So thank you very much. I hope these thoughts are valuable as we look at 2024.

Okay. So we've heard a ton about generative AI from all of you. This is our time now with all 4 of the analysts to do a little bit of a Q&A and talk about some of the things maybe we didn't touch on or we touched on very quickly. I feel like we just scratched the surface of so many topics. I'm going to post some questions here and then just also a reminder, keep sending in your questions through the platform and we will have time to began to take those as well. So George just mentioned in his talk about trust, I think you all hit on trust at some point in extent. And we probably very appropriately set up with the urgency of you've got to act now and, "Oh my gosh, you can't lose trust because trust is really important." So I think it's worth just doubling back on this for a second and make sure we're clear about it. Audrey, I'm going to start with you. Can you give us a sense of the state of trust among consumers today and then why it's so important as folks move forward with gen AI?

Yes. It definitely is the underlying theme, I feel like across predictions we talked a little bit about today. And for consumers, building trust, oftentimes, a lot of companies think it's a one or done thing, right? It's a reaction to things. But specifically with a very complicated, heavily debated technology like generative AI, it's an ongoing thing. It needs to be a metric that's constantly measured and reported on. And in my prediction earlier, I talked a lot about skepticism. And it will still be there. Skepticism about the product, skepticism behind the morality and the ethics of it. But our data has also shown that the more use cases the consumer has under their belt with generative AI, the more likely they will trust it. So a lot of the distrust for consumers actually happens because they don't know or it's in the unknown. So it's really important in the future as companies build applications and technologies and platforms with generative AI, it's really important to make sure that they build that trust so that consumers are more comfortable with using it and making sure they're getting it right.

And Laura, could you -- is it similar with business buyers? Or are they taking a little bit of a different tack, different attitudes what would you add there?

So when it comes to business buyers, Sharyn, we at Forrester, have been looking at this through the lens of the 7 levers of trust. And this is a research that we've been doing for a number of years. And that research, even our most recent in 2023 shows that trust directly drives positive business outcomes. So it is important and it directly impacts your business. So our research shows, for example, that business buyers who trust a company are twice as likely to recommend them to somebody else. So as we track through the 7 levels of trust, what we found is that buyers value all 7 highly, but competence, consistency and dependability matter most today. Now I think this has interesting implications for using generative AI because buyers say they trust vendors who are competent, consistent and dependable. And if you lose -- if you lose AI-generated stuff on them that make you appear to be less expert to act inconsistently or not in a way that they're predicting you to act or in ways that meet their needs, those very directly are going to erode that trust. So I think everybody has probably mentioned, I know Martin and Audrey for sure that it's important to make sure you put reviews and checks and balances within your -- all of your processes, whether it's customer experience design, B2B marketing, sales, product design. So that it's happening consistently and it's minimizing as much as possible any poor experiences, any inconsistencies, any mistakes that happen.

Okay. So trust is super important, clearly. We want to make sure we're minimizing anything that would erode trust. If I was a super risk-adverse person, I would say, well, then I'm not touching that stuff, right? This seems scary, I could erode trust very easily because there's a lot of things that could happen that I don't even know. It's brand new. And yet at the same time, we're saying, move really fast. You've got to move. You can't wait, et cetera. Martin, I'm going to put you on the spot a little bit. Can you fail? Is it okay to experiment with gen AI and fail even if it's going to erode trust a little bit?

I think the key to the answer is in the words you used in the question, which is experiment and fail. If you are not failing at probably, what, [indiscernible] experiments, you're not experimenting, you're just doing stuff you already know how to do. So the whole point of fail forward test alone, all those kind of activities is to fail but to fail in a very controlled way. So absolutely fail with your UX fail with whether it drives the right business case, fail whether it even works and delivers what we want, don't fail on bias, on privacy on the things we talked about earlier on, legislated things in legislative markets. There are ways to do that with like ring-fenced audiences, with test customers lean into internal use cases. We've all talked about analyzing data. We're helping you create content, connecting internal use cases. And I think that -- did you know that [indiscernible] word of the year was authenticity. So I think authenticity trust goes together in terms of build things that feel transparent, robust the customers want, but I don't know just control the scope is the best advice I could give, but absolutely don't be afraid to fail.

Yes, control the scope, and it sounds like also along the lines of that authenticity, be transparent. If you're experimenting, especially for experimenting and with a customer-facing use case. I don't know it's an experiment. So that...

Absolutely. The data -- I think -- yes, myself and Audrey, both presented data that suggests like 2/3 or more of customers want you to be transparent and be really open. And actually, you're okay when you do that. And we say out of our trust model, transparency is trust superpower. So that's be transparent.

Perfect.

Sharyn, just real quick, I think that bias towards experimentation is also what's behind George's recommendation that people start 10 projects. Knowing that some of them aren't going to go much further or aren't going to be successful?

It's a very good point, something so new. We've got to experiment and be okay to fail with some of them. Okay, use cases. I'm actually already seeing questions coming in from folks in the audience about this. So can each of you give me and I'll let you guys decide who wants to start. Give me, I don't know, 1 or 2 very specific use cases where there's been proven ROI or just -- it's worked really well versus been a failure it could be something really specific or a list of 3 that are just like boom, boom, boom these are very, very common. Yes, go ahead, Audrey.

Happy to kick us off on this. And I'll talk about it from a use case that I've seen over and over again consumer perspective and from what we've seen in our data, too, which is one is a big one, which is language translation. And I'm very obsessed with understanding what are -- what is going to get skeptical consumers, what is going to kind of overlay that hesitancy for consumers. And language translation is a capability that I've seen over and over again, providing functional utility to consumers. So in our data, we've seen it be the one function that the group who we see are more cautious around generative AI and AI with that technology, believe to be more useful. So that's definitely something that I'm keeping my eye on to.