Veritone, Inc. (VERI) Earnings Call Transcript & Summary

Now I'd like to invite Victoria to walk through the products that are driving the momentum and show you exactly how agencies are using Veritone technology to transform public safety operations. Victoria take it away.

Thanks, Jon. I'm Victoria Dickson, and I'm the Product Director for our Public Sector Solutions. Over the last 4 years at Veritone, I've led mission-critical initiatives across multiple businesses, increasing customer value, internal adoption and more efficient go-to-market execution. Starting with investigator. Investigate is one workspace for every artifact in a case, including body cam footage, CCTV, interviews, documents and more. Let's open an example of an officer involve shooting case. The detective can open the officer's body worn camera, read through the summary generated on upload to quickly gain context without watching the whole clip. Now on the time line, she can search the transcript, objects and faces in one view and jump to key moments in seconds, not ours. Switching to detective, searching across sources who need to connect the dots. They can find a silver minivan and CCTV footage and on social media, if they've shared in the post the next day, all inside of one system. And for all cases and cross case searches, we ensured officer protections are in place. We provide incentive content alerts for relevant files based on key file intelligence gained upon upload. By default, this covers CCM and pornography, but is fully customizable based on an agency's needs all available today, one pane of glass, intelligence on upload, cross case searching and investigatory protection, all built in. Now for what's next, nested folders and map-based preview. Soon, you can organize your cases as needed, segmenting out sources or isolating key evidence. The upgraded key actions will give you speed for efficiency for items like tagging, editing and exporting. Additionally, the new map view will give you key geographic intelligence at every step of the way inside of your case as well as across cases, production powered today with key enhancements coming soon. Now let's move to Redact. Redact is one of our most widely used public sector applications and solves a very real challenge. Agencies are required to protect personally identifiable information when releasing body-worn camera footage, dash cam footage or other key evidence. Historically, this meant someone had to go frame by frame and manually blur sensitive content. Redact removes that burden. The application features object detection flags for our faces, heads, allows you to confirm redactions for other objects like laptops, notepads and license plates. You can also manually confirm any of the redactions that are auto detected. You can apply codes or leave timestamp comments for others to review. For audio on the same file or for audio-only files, you also have the ability to come in and redact or apply redact voice masks. All depending on your agency's needs and file release policy needs. Images follow the same detection review workflows as videos without displaying the time line video or audio tabs. Video, audio and newly released image redaction capabilities along with audit reports, all available in production today. And now a preview of the soon-to-be released document redaction. For PII detection group or addresses, phone numbers and names, you can now check and confirm the findings. It allows human review and release instead of audio release, outlying preview also shows what's behind the box, so you can scan through and review to make sure everything that you need is covered. Redacted files produce the same output, which is a redacted file with audit capabilities ready for council review or release to the requester. Once the document redaction capabilities are released, you'll have access to redact all media in easy-to-use audit friendly tool. And finally, I'm proud to show you, Assess, which is a brand-new AI-powered data analyst solution, but with law enforcement, military and compliance professionals in mind, Assess condenses time spent on manual data reviews, empowering investigations and ultimately transforming unstructured data and actionable insights. Currently available as a stand-alone application, this tool will soon be available within investigate. The platform features three primary capabilities, shipping together. Persona configured agents, Barry chat and formal case assessment. Agencies can customize and define specific personas, such as homicide, narcotics or financial crimes within the settings. Within the case workspace, investigators can use plain language to [ Asbury ] similar to what Alyssa previously demoed as part of our commercial solutions suite to perform tasks, such as finding every case tied to a specific phone number, [ Veritone ] scans all key data stored across files and cases to provide actionable guidance. For instance, we can filter 12 active cases in one step, surfacing patterns in minutes that would typically have taken days to cross reference. The case assessment feature analyzes incident reports against department policies for command and internal affairs review. Time line findings are easily categorized with each link directly to the source document. This allows reviewers to validate the reasoning behind findings rather than just the conclusion. Finally, reports can be easily exerted from the same interface, providing structured and auditable information for leadership. The platform integrates daily investigation tools with cases for oversight, creating a single platform that handles everything from evidence to briefing and release. Thank you for your time today. Before I turn it back over to Peter to share more about our data licensing business, let's take a minute to hear from one of our partners, Oracle.

When people talk about AI, they often focus on the models. What we're seeing with companies like Veritone is that the real opportunity lies and what organizations can do with their data. Across media, entertainment and sports, companies are sitting on decade valuable content. Video, audio images and documents that has historically been difficult to access, analyze and monetize its scale. AI is changing that. Veritone is helping some of the world's most recognizable media brands unlock the value of those assets through intelligent workflows that make content searchable, actionable and revenue generating. That started the possible when AI is paired with the right data foundation. As those workloads grow, scale matters, the amount of compute required to process, analyze and derive intelligence from petabytes of content is substantial and it's only increasing as AI adoption accelerates. That's where Oracle comes in. Oracle Cloud Infrastructure provides the performance, reliability and scalability needed to support these mission-critical AI workloads. Together, Oracle and Veritone are creating an environment where customers can confidently apply AI across massive content libraries and operational workflows without compromising performance or trust. We're excited about the partnership because it represents more than a technology deployment. It's an example of how AI can create entirely new opportunities for media, entertainment and sports organizations, and we're still in the early innings of what's possible.

The AI narrative has fundamentally shifted from raw computing power to high-fidelity data architecture. Data is the single most critical element of the entire technology stack. For years, early foundational models relied on scraping public internet data. Today, that public resource has been exhausted. The true competitive advantage now belongs exclusively to those who possess unique, proprietary and recurring real-world information. Enterprise data is the ultimate competitive moat and fuel behind foundational models, application models or internal models. However, a massive commercial bottleneck remains. The world's most valuable data is multimodal. Hundreds of millions of hours of rich video and audio trapped inside unstructured legacy systems and archives. Because models are only as good as the data used to train them, AI researchers, scientists, and machine learning engineers are now allocating billions of dollars to secure premium ethically sourced and rights cleared datasets. Veritone sits precisely at the intersection of these market forces, transforming dormant enterprise data into a highly liquid, entirely new asset class to monetize enterprise data, you must first translate it into a universal language that machines can comprehend. That is the function of the Veritone data refinery or VDR. Think of VDR as the industrial conversion engine that transforms raw unstructured media into AI-ready assets. Organizations sit on decades of audio, video, images and documents that are historically impossible to index search or utilize. VDR ingests enriches annotates and formats this dark data into high-density machine-readable tokens. Crucially, and when necessary, the platform can also execute automated privacy and compliance controls, such as the redaction of personally identifiable information. Veritone monetizes VDR through a highly scalable transaction matrix including data processing fees, consumption-based refinement revenue and data preparation services. The global traction is undeniable. The Veritone data refinery processed 22.2 trillion tokens during the second half of 2025 alone. This marks an over 3.5x volume increase compared to the first half of the year, driven directly by the throughput demands of all the leading public hyperscalers and model developers. This extraordinary momentum is rooted in deep institutional trust. As Sean mentioned, we have over a decade of experience connecting rights holders with buyers through our content licensing business. Our content and data licensing experts have executed thousands of agreements across media sports news and entertainment for some of the world's most recognized brands. Our experience negotiating usage rights, managing approvals and helping support compliance through evolving copyright and distribution standards -- in addition to our established reputation for securely facilitating massive volumes of rich multimodal data through our platform has solidified our core partnerships with global hyperscalers, consistently yielding multiple repeat orders and extensional benchmarking validation under the most demanding enterprise workloads. If the Veritone data refinery is the industrial plant that processes raw content into high-grade fuel, the Veritone Data Marketplace or VDM is the financial marketplace where that fuel is securely transacted. VDR produces the asset and VDM monetizes it. VDM is a fully governed, transparent marketplace built to establish clean provenance and fair compensation. On one side house of the exact same macroeconomic by positioning Veritone urging data economy, we are empowering organizations to unlock the true liquidity of their archives while securely provisioning the next generation of artificial intelligence. Maximize your data, empower the future, this is Veritone. With that, I'm pleased to welcome Sean King back to share more about our talent acquisition business, Broadbean by Veritone.

Thanks, Peter. While much of today's discussion has been focused on AI and digital transformation, Broadbean represents something uniquely important to our long-term strategy. It is a durable global SaaS business with deep customer relationships, highly recurring revenue and significant opportunities for innovation. For more than 2 decades, Broadbean has helped employers solve one of the most fundamental challenges facing every organization. finding and hiring talent. Today, Broadbean serves thousands of customers around the world. We're talking about enterprise employers, government agencies, staffing organizations and recruiting across virtually every industry. Our platform helps them distribute jobs, manage recruitment and marketing campaigns, optimize candidate attraction and improve hiring outcomes. At a time when organizations face intense pressure to do more with fewer resources, these capabilities have never been more important. The recruiting landscape has become dramatically more complex. Employers are expecting to attract talent across dozens of different channels. candidates expect faster personalized experiences and recruiters are being asked to manage crushing workloads while approving measurable results all while hiring budgets remain under pressure. This is why our business model is so incredibly resilient. We generate revenue through reoccurring SaaS subscriptions, recruitment advertising services and strategic technology partnerships because recruiting remains a mission-critical function regardless of the economic condition. Broadbean benefits from long-standing customer relationships and highly predictable usage patterns. Many of our customers have relied on Broadbean for years. and, in some cases, decades. As a foundational part of their hiring technology stack. That level of durability is incredibly valuable to Veritone. It provides a stable source of recurring revenue while creating built-in opportunities to expand services, introduce new capabilities and deepen customer relationships over time. One of the primary reasons Broadbean has maintained its market leadership is its ability to integrate seamlessly into the systems customers already use. Today's hiring environment isn't powered by a single platform. It's a vast ecosystem of applicant tracking systems, job boards and analytic tools. Broadbean acts as the connective tissue across that entire ecosystem. Our deep integrations with industry leaders like Workday, Oracle and SAP position us right at the center of some of the world's largest hiring environments. These relationships create a massive competitive moat that give us a distinct advantage and provide immediate access to growth opportunities as organizations continue to modernize their talent acquisition. Together, the Broadbean suite includes four core solution areas designed to maximize operational efficiency. First is Broadbean job distribution. This enables recruiters to publish openings across hundreds of job boards and career sites from a single interface. Instead of manually posting jobs one by one, recruiters distribute positions at scale, maintain consistency and instant visibility into performance. It creates massive operational efficiencies. Second is Broadbean programmatic. As recruiting becomes increasingly data-driven, organizations need smarter ways to allocate ad spend. Programmatic helps employers automatically optimize recruitment advertising using real-time performance data. Instead of guessing where to invest recruitment dollars automation, direct spend towards the channels delivering the strongest results. This means higher candidate quality, lower acquisition costs and a much better return on investment. Our recently introduced job acceleration capabilities extend this value even further, allowing recruiters to quickly boost visibility for critical or hard-to-fill positions when hiring needs are most urgent. Third is Broadbean OFCCP. Compliance is a strict critical requirement for federal contractors and highly regulated organizations. This solution automates and simplifies complex compliant workflows reducing the administrative burden while mitigating risk and maintaining confidence in the hiring process. The fourth is Broadbean Media Services. This is our managed services offering that helps organizations maximize the recruitment marketing investments. Our team works directly with customers to develop media strategies, negotiate placements and improve outcomes. Media Services continues to demonstrate strong growth as organizations increasingly seek expert guidance to navigate a fragmented marketplace. It adds a powerful layer of value that drives long-term retention. But what makes Broadbean particularly exciting today is not just where the business has been, it's where the business is going. Recruiters spend countless hours performing repetitive administrative tasks, writing job descriptions, managing campaigns, analyzing data, coordinating workflows. These processes are the exact use case that benefit most from intelligent automation and agentic AI. This is why we are investing heavily in the next generation of Broadbean capabilities. We believe the future of talent acquisition isn't simply about posting jobs. It's about creating intelligent systems that help recruiters make better decisions, move faster and focus their time on high-value human interactions. Broadbean is uniquely positioned to lead this transition because we already sit at the center of the workflow. We have the data, we have the integrations and we have the customer relationships necessary to bring these innovations to market at scale. As we look ahead, we see significant opportunities to expand Broadbean's value positioning through automation, AI assisted recruiting and deeper HR ecosystem integrations, combined with its recurring revenue model, global customer base, and powerhouse partner network, Broadbean remains an important profitable contributor to Veritone's long-term growth strategy. It is a business perfectly positioned to ride the very same AI tailwinds that are reshaping industries all around the world to show you exactly how these innovations are coming to life. I'd like to turn it over to Brittany who will give you a look into what the future of Talent acquisition looks like with Broadbean. Brittany, over to you.

Thanks, Sean. I'm Brittany McKenzie, the Senior Director of Sales and Account Management here at Broadbean by Veritone. There's always that one job, the one that hiring team needs filled right now, and it's sitting in the same campaign as everything else, taking the same slice of budget as roles that aren't urgent at all. That's the gap job acceleration closes. It takes a single role and gives its own dedicated push without touching anything else you're running. Let me walk you through it. Every job in the campaign has its own accelerate control. right here in the row. Opening it brings it up to a quick setup just for the one role. You're choosing a budget, a run time and how quickly you want candidates coming in. Nothing else in the account is affected. And the whole thing rests on three simple ideas. The budget is dedicated, so it never pulls from your main campaign. The delivery is faster because this role now has its own push behind it, and you only pay for the traffic you actually use. There's no waste built in it. So when a budget is chosen here, it funds a brand-new separate campaign for this one job. The original budget stays exactly where it was. You're adding firepower to the role that needs it, not borrowing it from somewhere else. One confirmation in the new campaign is cleared to run right alongside the original job. From there, it goes live straight away. It starts competing for the candidates on this role immediately. with no waiting for the next cycle to come around. And that's it. The role is accelerated. There's a new campaign running now, tracked entirely on its own. So the impact of that extra push is clear from day 1. Back in the table, every accelerated role carries a badge. So at a glance, even across hundreds of jobs, you can see exactly which ones have a dedicated push behind them. Open the card and you get the live picture for the role in one place, clicks, applicants, budget, spend, cost per click, cost per applicant, everything you need to know that it's working. And you stay in control the whole way through. The same card lets you end the acceleration whenever you want. If priority shift or the role gets filled, you stop the spend right there. Now here's the part that really matters. The applicants from the accelerated role sit right alongside every other job in the table. So you can see in plain numbers, the candidates, the dedicated push brought in. That's the return in black and white. And all of this rolls up, whether you're in the overview, the jobs view or reports, the acceleration data comes through. So you can compare its impact across the whole account, not just one role at a time. So here's what I'd leave you with. When a single role needs to move now, job acceleration gives it a dedicated push, measured on its own with spend you stay in control of. One of our customers, SOS group said it better than I can in their words. This has been an absolute game changer for us. We have so many times needed to be reactive and here it is. This is how we can do that. This is the reaction we keep hearing, and it's exactly why we built this.

And now let's take a moment to hear from one of our customers.

My name is [ Megan Gabi ], I am the Director of Talent Acquisition with SOS Group. Prior to working with Broadbean, we were experiencing a lot of issues with getting a large pool of applicants. So the volume of applicants we were getting was just not sufficient enough to fill the amount of job openings that we had. And we needed to address this in order to please our clients. The clients are -- we're their vendors, we work for them, and we want to make them happy and the jobs were not getting filled. And even though there may have been a few jobs being filled, they weren't being filled fast enough. So we engaged with Broadbean to help us bring in just a larger volume of applicants. -- what we were doing wasn't enough. We were posting manually. We were updating our own post. We were revamping our own post on one job board, one of the largest job boards out there that everybody else was using. So the competition was fierce, and we couldn't keep up. So the Broadbean solution that we used, I mean, just using the sponsored post getting us out there on top of multiple job boards, we are using the sponsored posts through Broadbean. It's been incredible. Time line is pretty quick. I mean it was a very seamless process once we got involved with them and the integration process for us, everything was very straightforward. And I mean, kind of hands off for us as well. So Broadbean and Veritone is who we're currently working with. They were able to just jump into everything that we needed with minimal need from us, they were able to get it all done. It honestly did not take long at all. We ended up probably seeing the turnaround within just a day or 2 of our jobs being put on boards and applicants rolling in. We've seen quite a few benefits from this. I mean, again, just seeing a larger pool of applicants there's a lot of competition out there right now. And we have just seen a tremendous amount of more qualified applicants, just better fits rolling into our system. We now have a database filled with all types of qualified applicants across the board. Since we started using Veritone and Broadbean, we have really noticed the number of -- some of our KPIs were -- that improved where the number of interviews the recruiters were getting weekly, those -- some weeks have tripled with the amount of applicants we're pulling in, especially on our busy time of year, which is now. We have seen a tremendous number of official her letters, offer letters going out. So from the funnel of recruiting from that net we cast out -- we have seen a huge increase in that. And even though at the bottom of that funnel, it doesn't look like a huge number and the number that we do get coming out and working for us is much larger than it ever was prior to us using these services. One of the most important data points I would consider would probably just be, again, the number of applicants we get. Staffing, there's no exact science to it. There's algorithms out there that help us get people, but they're still very unpredictable. Working with the human race is just very challenging at times. But that being said, we've learned anything. It's the more people we talk to, the more applicants we get, the higher chance we get of somebody making it across the finish line and starting to work. So we -- I would say the number of interviews increasing and then the number of hire letters were just so important to us and have made a huge difference. It's helped our company grow over a short period of time. And then learning a little bit about the job acceleration, too, we are now able to become very reactive when needed. Our clients have noticed and our clients have made positive comments on us filling jobs.

All right. Great. Sorry for that delay as we transitioned over. Anyway, thank you for you guys for staying on. It was a long session, but obviously, I think -- the point where we want to really impress upon everybody is, well, first and foremost, I thank -- want to thank my team for putting this together in all of our customers are really and eagerly willing to speak out about what our solutions are bringing to the table and how they're part of not just their day in and day out and weekly and monthly workflows, but they are mission-critical as you've heard from them and you've heard from our team. What you've seen and what we are building in the point, the important elements here are, it's all about data and AI, right? The magic and what you can do and how you can apply orchestrated AI and agentic AI to massive amounts of messy unstructured data with real results and obviously, we've been able to do that across multiple business segments and in some areas, radical different workflows and use cases. But the important element is that these are all built on a common vision, a common stack of aiWARE and it allows us to continue to service these customers at huge scale with great velocity. Also, the game is changing. Our need to continue to innovate fast, introduce the latest state-of-the-art features and functionality is critical for us to stay competitive, and we are. So again, I want to appreciate everything -- everybody for joining today. We have been compiling a list of different questions. I want to spend a few minutes who those can give me the audience to stay on for a little bit, and I'll do it. Please note that this presentation is going to be made available for replay. Obviously, there's a lot of great nuggets in here that we're going to intend to check up and distribute through different campaigns and the like. But again, this entire webcast will be available for playback. But let's get to a few questions. In terms of scale, one thing I wanted to add before I jump to Oracle is what you've seen here today and what we're building is used by tens of thousands of unique users and thousands of customers each and every month consistently. So again, we just want to double tap on what you're seeing here our mission-critical workflows used by end users and enterprises every single day, every single week and every single month. And again, that's something that we take a lot of pride in, but it's important that we continue to build and extend upon that opportunity.

All right. For the first question is about Oracle. I'm just going to give you a question was specifically, can you provide a quick update on Oracle and where we are with that product integration, the product rollout and the extent of the relationship. We are on track. And I think we are planning on right now to start moving a material amount of data payloads initially focused on the data side. And then secondarily, on the compute side as early as August, which is right on track. More importantly than just a phenomenal technology partner, we do expect to be clear when we make this transition that the efficiencies that we're going to realize through this relationship could potentially generate up to 20% or more in cost savings, which is going to be just awesome. But however, there's a lot more to this relationship and then just being a technology vendor. We look at this as a great partnership for co-selling and really attacking the market as really like partners, almost like teaming agreements. I had the opportunity to speak live at the annual sales kickoff to an audience of thousands of sellers. And frankly, it was just Veritone and one other customer who had that opportunity. So again, we're -- again, we are all in on the Oracle relationship. They've been a great partner so far, both in helping us assist in the transition on the technical and the build on the technical side, but also a co-selling partner as well. All right. Second question is a little bit on just what's the business model or a little bit of the workflow behind VDR. We got a lot of questions here, but I'm going to kind of give you short and succinct answers here on some of these. VDR is exciting, and we sit on and we represent -- we said on both sides of the equation, both the buy side and the sell side. On the buy side, we have direct contracted engagements with numerous foundational model in hyperscaler companies, the likes of the Googles, the NVIDIAs and others. So again, we're under a master services agreement contracts with them, and then we execute specific task orders or purchase orders as data needs come out. So for example, if they're looking for 10,000 hours of security push that needs to be run through aiWARE, indexed with metadata, cleans, transcoded, et cetera. And that would be an example where we're executing an order working with those hyperscalers directly. On the flip side, we also have numerous engagements and we are under contract on the sell side. So the likes of a lot of the data suppliers. So again, the vision here is to have the largest supply available, readily available. So we are always looking to acquire new data sets, build relationship with new data providers and immediately through our technology stack, DMH and aiWARE is make those data sets available and get those supplies and those samples in front of these data scientists in front of these hyperscalers as quickly as possible. Again, we think about it as hit rate. We don't want to be in a position when a hyperscaler says, "Hey, we are interested or we're looking for a data set, how quickly can we find it if we don't have it already and how quickly can we prepare that data set, process it, enrich it with metadata and be able to prepare it and deliver it to the end customer. So again, VDR, Veritone Data Refinery, we sit on both sides, contracted with the buyers and the seller side. It's a very advantageous position that we're in today. A question about hyperscalers and how we look at them as partners or competitors. We look at them as primarily partners. I could say, Oracle, obviously, is a hyperscaler on infrastructure side, AWS, NVIDIA, these are obviously hyperscalers, and they're both partners, right, where we are actually selling -- they're actually generating revenue from them, but in many instances, are co-selling with them in the likes of an Oracle or an AWS. Of course, they offer their own models. And of course, they offer their own application frameworks where you can build on the solutions. But again, what makes Veritone so powerful is we are the full end-to-end stack right, from the full in data ingest, the orchestration of models, right, that we're agnostic, so people can bring their models or they can tap into over 250 models that we have readily available on our stack. All the way to using our API or application levels. So again, Veritone can be the full end-to-end stack or we can work in concert in coordination with the hyperscalers, which we've done on multiple occasions for different customers out there. So again, the ecosystem, obviously, there is overlap in different areas, but I would definitively state that a lot of these groups are either clients or partners, not necessarily competitors of ours. Let's talk about highway patrol. Recent large deal that we just announced around our public safety software solutions, specifically Redaction was with the California Highway Patrol. It was a multiyear deal servicing tens of thousands of officers. Yes, we look to be able to repeat that. Now first and foremost, the products that we've sold to them are rinse and repeat. These are highly reliable, highly available software solutions that can immediately be provisioned and spun up for new customers. Yes, we are always -- as we kind of mentioned and Jon Gacek through his presentation today, and did mention, Veritone continues to penetrate larger and larger agencies out there. And that's the whole game is fight hard to get those initial wins, overservice the companies with great products and great services and great service. and continue to grow and expand there. But absolutely, we're thrilled about California Highway Patrol, which is a very large agency. We're looking to expand, not frankly just within that agency, but rents and repeat with our multiple different solution offerings and be able to take those to multiple different agencies across the country. Broadbean by Veritone. We talked a lot about that at the very end. Broadbean just is a phenomenal business. It's very profitable. It's got a very large customer base. But to be clear, like the rest of our business offerings, it's a data problem. They are helping solve massive data and data coordination and workflow problems. They are currently integrated with hundreds of different ATS, application tracking systems like the Workdays and the Oracles out there, but they're also integrated with thousands of job boards and their business model is really helping businesses who are looking to hire individuals be able to sift through all those job recs, find the right job-based advertising and distribute those to thousands of job boards. And then process millions and millions of applicants each and every single month. So it's a massive data workflow. It's a massive data problem, ripe for obviously how we're applying AI to make that process ever more efficient. We love the business. It's got a very large international customer base and stay tuned. We're going to be rolling out a lot of new features and net new products and applications through the course of the balance of the year. going after them. And also cross-selling. There's -- Broadbean has hundreds already of public safety customers both domestically and internationally. So again, cross-selling our products and solutions across our expanded customer base remains a very important priority for us with clear OKRs. And again, we've already started to do that, and broadband is going to be a big part of that cross-selling and growth strategy. Let's do it just a couple more. Back to some of the reorganization stuff that we've been talking a little bit on previous finance calls and earnings calls. We did announce that we are going through major restructuring and really with a concerted effort to reduce our cost structure. As I can report that we are on track, I think we were very clear that we are looking to execute a lot of these reorg and cost-saving initiatives between June and through the end of July, we are on track and we're going to stay disciplined and we're going to get it done. Again, this is a business that has a great opportunity. It's got a very large insole revenue base, thousands of customers, but we've got to rightsize our cost structure. Veritone obviously should be in a different position. We're obviously not happy with the way that the stock has been performing for those from the investors who are listening to this. But as everybody's seen here today, we have great technology, state-of-the-art technology. We have thousands of customers, very engaged customers. And we got to fix -- and we have been, but we've got to fix the other areas of the business. As we touched on before, we've deleveraged the business significantly coming from an aggregated total debt load of well over $200 million in the past. We've brought that down significantly. But again, there's more work to be done, and we're going to get it done. So again, I want to leave here is, first, what you've seen today is awesome. It's brilliant and exciting. I'm thankful for my team. I'm thankful for the customers who again, lent their voice to support us. And this is a very passionate group that's running this business top of the bottom and we're going to get it done and Veritone is going to be a stalwart in the AI and data ecosystem for the years to come. So with that, I'm going to leave sign off here. Again, this will be available for replay later and we're also going to be leveraging a lot of this great content and chunk it up into different campaigns as well. Thank you for your time, and we'll talk to you soon. Bye-bye.