Catenai (LSE:CTAI) — Investor Presentation

0:00 Good morning, and welcome to the CATENAI PLC investor presentation. Throughout this quarter presentation, investors will be in listen only mode. Questions are encouraged and can be submitted at any time by the Q and A tab situated at the right hand corner of your screen. Just simply type in your questions and press send. The company may not be in a position to answer every question it receives in the meeting itself. Have the company can review all the questions submitted today and publish responses where it's appropriate to do so. Before we begin, I'd like to make the following poll. I'd now like to hand you over to interim CEO, John Farthing. Good morning to you, sir. Good morning. Hello, everybody, and thank you very much for joining us this morning. I'm delighted to introduce Iludium, which is our primary investment as shareholders will know. And they've made a lot of progress over the summer, so the objective of today's presentation is for them to showcase what they've been up to and what you might look forward to in the future.

0:49 So we're delighted to be joined by John Frizzell, the chief executive of Iludium, and Dave Burrows, the chief commercial officer of Iludium. And with that, I'll hand over to John Frizzell. Thank you. Thank you, John. Good morning, folks. So today, myself and Dave are going to be walking you through a presentation and a demo of Illudium, setting the stage for, excuse me, the AI market and AI agents in particular. So let's start with a little bit of background, a little bit of context to get us started. So what we've seen over the last couple of years is the ChatGPT has become one of the fastest growing products in history, both for consumers and organizations. And it's not just their product, OpenAI, we've also seen other LLM vendors such as Google and Tropic, DeepSeek, and Mistral commanding exceptionally high valuations.

1:42 Now these foundational models have become incredibly powerful, but in reality, organizations are only using a fraction of their potential because they're still primarily being used as chatbots or copilots. To get the real benefit from AI, organizations need to use AI agents. Agents are the next big wave of AI, as big if not bigger than what we've seen with Alexa ChatGPT, Google Gemini, or Cursor, which we'll talk about next. So we have built and are launching the next generation AI agent platform. It allows agents to be rapidly and easily built through a simple chat interface and used by individuals or teams via a workspace and chat app. Our vision is that every person needs a digital team of agents.

2:30 We believe that Illudium has the potential to be as big, if not bigger, than companies like Cursor, which is a developer tool, or Lovable, which is a tool for generating user interfaces. These are both billion dollar companies as we can see on the right hand side, as well as companies like Perplexity and Eleven Labs, which is a text to voice and voice to text company. The team behind Illudium have experience doing this. I was a founder of a company called Feed Henry that we exited to Red Hat back in 2014. That was a mobile enterprise application company. We're building something quite similar with Iluvium. It is an enterprise AI agent platform, and we believe that it has massive, massive potential.

3:17 And I'll hand you over to Dave to take us through some of the positioning. Thanks, John. So we've had a lot of discussion in AI around content creation, chat GPT, copilots, the whole first wave. Then we've moved into reasoning models, increasingly capable, increasingly powerful. But the third wave is different. It's not about creating content or reasoning about problems. It's about doing the work, and that's where Iluium comes in. At a very simple level, we're building two tools. One tool is for people who want to build agents, and they can do that simply by chatting with the builder. And the second tool is for people who want to use agents.

4:02 No complexity, no setup. Engaging with the ludium is as simple as using WhatsApp, and John's gonna walk you through this in the demo. Our vision is a future where every person has a team of agents working for them. Humans set the direction. Agents handle the execution. They take care of the repetitive operational work so people can focus on judgment, creativity, and decision making. But this is not about replacing jobs. It's about elevating human productivity. One of the easiest ways to think about agents is as digital coworkers. In this case, Elena.

4:47 Elena is a VC type analyst as a digital coworker. She has areas of expertise such as sourcing research and portfolio support. She also has real tools she can use, such as LinkedIn, web search, internal documents, and CRM style data. Where this gets really interesting is how that turns into applications. So for example, Eleanor can use LinkedIn and web search to monitor founders and companies, flagging up relevant updates automatically. She can use documents and internal data to help prepare research, diligence notes, or IC memos. And she can keep running in the background, supporting portfolio work over time and as a one off task. Thinking about agents this way, specialized with defined skills, tools, and applications, makes the whole concept much more intuitive.

5:40 Once you have Eleanor, our digital coworker, the next step is obvious. You start grouping them into teams. Teams of agents working together on your behalf, coordinated, specialized, and always available. Ultimately, this leads to entire organizations of digital coworkers. On day one, what you're getting is Elna and multiple versions of her. Teams come later, and organizational automation follows that, supercharging organizational productivity. Let's hand back to John. So as Dave said, we think about AI agents as digital coworkers that can supercharge productivity.

6:26 When these agents are set up correctly and put to work on the right tasks with human oversight, what we see is that these digital workers can outperform human workers on certain categories of work. Whether it's availability with these agents being available twenty four seven, consistency and focus, quality and reliability, cost, collaboration, scalability, speed of work, or adaptability to change. These digital coworkers are positioned to really automate a vast amount of work quicker, cheaper, and more efficiently than humans are currently doing. So the term agents is very much in the zeitgeist at the moment. But as with humans and human roles, there is not just one type of agent.

7:16 So we look at agents through two different lenses, the functional roles they take and the level of autonomy they have. Within functional roles, we first have task agents. These are small focused agents that complete well defined tasks end to end. They're on a narrow loop, they get work done. Above that, we have workflow agents. These are coordinators. They chain tasks together, delegate work to the task agents, and manage processes. We then have domain expert agents. These are reasoners. These use the reasoning models like o one and o three. They do deep reasoning deep reasoning within a specific knowledge domain, preparing reports, and producing detailed domain expert analysis.

8:04 And finally, we have interface agents. These are right at the surface. These are the front door between humans and the agent ecosystem. These are the agents that humans interact directly with and task them or their colleagues will work. When we think about autonomy levels, what we see is first reactive agents. This is what most people would be familiar with. This is an agent that doesn't run on its own. It doesn't run-in the background. It responds only when prompted. This is the experience people will be familiar with from the likes of chat GPT or Claude. Then we have planning agents. These are agents that break large bodies of work into smaller steps, choose what tools they're going to use, and manage multi step tasks and workflows.

8:52 We then have autonomous agents. These are self initiating agents. They run-in the background, triggered either on a schedule or by external events. They work on a set of tasks, and then they reach out to the human they're collaborating with to bring them back into the conversation at exactly the right time with updates, results, or queries. And when you bring all of that together, what you get is these multi agent systems. Teams of agents coordinating with each other using delegation and shared context to ensure that they're all working in the right way on the right tasks. Regardless of the type of agent, what we have identified is that there is a common set of needs that agents have, and this is what we have built, the operating system for agent execution.

9:47 So looking at the diagram on the right hand side, at the top, we have the integration layer. This is what allows agents to connect, connect to external systems, connect to document repositories, and connect to other agents. We then have the intelligence layer. This is what allows agents to understand. They use large language models of their brain for thinking and reasoning. They generate and use memories, so they learn on the job and create new memories that they can refer back to. And they use context, information provided to them that grounds them in the work they're doing. Then we have the execution there. This is where agents actually plan and take action. And whether it's an autonomous agent that is operated by triggers and events, or an agent that is collaborating with a human in the loop, in order to act, in order to do real work, these agents use tools that connect them to external systems, whether it's your email, databases, customer relationship management systems, document stores.

10:52 These agents use tools to perform real world work. Finally, underpinning all of that, we have the governance there. This is where we set up things like guardrails, observability, and policies to ensure that the agents are operating within the parameters that have been specified and performing the work that they have been tasked with correctly and in a way that can be monitored and observed. And what this leads to really is an experience where work really starts to feel like it's working itself. So imagine a scenario where you open your laptop in the morning and you have a concise briefing from your executive assistant agent outlining the day ahead, intelligently shifting meetings to give you focus time, providing drafts of key emails, and giving you actionable insights.

11:45 Throughout the day, these agents reach out to you and suggest specific actions. I've drafted a response for this email. I can reschedule this meeting to give you more focus time. Then at the end of the day, you get a digest of what was accomplished, what's queued up for tomorrow, and what needs your attention. And simply with this one workflow, and this is just one type of agent that can be built within the platform, our expectation based on initial analysis is that users will reclaim one full day of focus per week. So we're now going to take a look at a demo video to see this in action. So if we could play the video, please. Welcome to Alludium, the agent operating system for human AI collaboration.

12:36 With Alludium, creating dedicated AI agents is as simple as choosing from a catalog of pre built templates, covering a wide variety of use cases, or creating your own agent from scratch. In this demo, we will create a new agent, a personal assistant for helping with email and calendar management. Configuring your agent is as simple as explaining what you want your new agent to do, using text or voice. Once you provide your requirements, the configurator agent will work with you to build your bespoke agent It will proactively suggest the right approach, confirming details with you step by step as it works With full access to the entire Iludium platform, the configurator agent is able to search for available applications and tools, assign them to your agent, and configure your agent's prompt, greeting, and other settings.

13:31 For those that wish to take a more hands on approach, it is possible to switch between guiding the agent configurator and directly editing your agent's configuration. You can manage your agent's prompt, AI model, tools, and connections, as well as creating and configuring triggers to allow your agent to work autonomously. For example, sending you a morning summary of meetings and important emails at a time you choose. You can test your agent directly from within the agent builder at any time, interacting with it in a sandbox test environment and iterating on the design until your agent is working exactly how you want Once you are satisfied with your agent, simply deploy it Your agent goes live instantly and once live, you can share your agent with anyone of any skill or experience level.

14:29 Users of your agent do not need to know anything about building agents. They get a dedicated chat app for interacting with their agents. This dedicated chat app is available on both desktop and mobile so your agents are available to you wherever you go These agents are not just reactive They can work in the background and reach out to you via email, web and mobile notification as needed to bring you back into the conversation or to deliver valuable information directly to you. Alludium, making the power of AI agents available to everyone. Okay. So what we saw in that demo was a particular type of agent being created.

15:22 In that case, it was a personal assistant agent. So it was using a large language model. It was using communication tools like Gmail. But, really, the power of Illudium is that it is positioned at the heart of an ecosystem of hundreds and thousands of applications. Whether it's project management tools, sales and CRM tools, social media, document storage, the ULUDION platform can be used to build any type of agent for any use case. So you might build a social media assistant agent that is connected to LinkedIn, x, and Reddit to help monitor social media activity, draft and publish new posts, or create a sales development agent that is connected to your CRM system, monitoring new leads coming into your CRM, pushing information, for example, into Slack to make people aware, and updating spreadsheets as needed.

16:24 So the market opportunity. Three massive ways are converging to create unprecedented market opportunity. Iluvium sits precisely at the intersection of workflow automation, agentic AI platforms, and productivity SaaS. Our addressable market is every knowledge worker, roughly a billion people globally. This opportunity isn't theoretical. It's the natural evolution of how work gets done. Illyddium's revenue model is very similar to the one that Lovable uses. This is the model that Lovable used to reach $100,000,000 ARR in eight months. The revenue model goes live in February, and we're already live with a small group of design partners, and we'll continue onboarding additional design partners and waitlist users between now and then.

17:17 Platforms enabling automation and agents are commanding multibillion dollar valuations even when agents are still bolted onto legacy systems. What this tells us is clear. The orchestration layer for intelligent work is being priced in already. Iluvium is being built natively for that future, not retrofitted, which is the why this opportunity is so compelling right now. So this brings us on to why this matters now. Individuals will be increasingly leading teams of agents. The companies that provide that orchestration layer for the future will define the next era of enterprise software. And investors who join now aren't buying a product.

18:03 They're backing a platform that will run the future of work. So just a little bit more about the people behind this. You've met myself and Dave today. We also have Barry Downs as our executive chair and DC Callahan, a veteran in the startup space, previous four time founder who is our CMO and go to market expert. On the right hand side, you can see the development team at a recent off-site that we had in Waterford, Ireland. And this team has been working very hard on building out this product. We're incredibly excited about where we're at and where we're going. And that is really it. That is our vision. You lead, your agents follow.

18:53 The infrastructure for the agentic age is being built right now, and Iluvium is ideally positioned to take advantage of that and really lean into this next wave. We wanted to leave plenty of time for questions, so that is our presentation. Thank you very much. That's great. Well, thank you very much for your presentation. Ladies and gentlemen, please do continue to submit your questions. You can do so just by using the q and a tab that's situated on the top right hand corner of your screen. But just while the company take a few moments for you the questions that have been submitted, I'd like to remind you that recording of this presentation along with a copy of the slides and the published Q and A can be accessed via your investor dashboards. As you can see, we have received questions throughout today's presentation. And, John, what I'll do is I'll hand over to you to run through the q and a, and I'll pick up from you at the end.

19:40 Thank you very much. So I think, let's start with when can we get access. So, I don't know. Dave, do you want to start on that? Sure. I can take that. So we're live right now with our first design partner, and we're taking additional partners, over the coming weeks and progressively taking people off the wait list, and we'll be in contact with people from the wait list in the coming weeks. And what we wanna do is learn about their their use cases so we can really ensure we're prioritizing access in the right areas. So in summary, January is gonna be when people are gonna be coming in, more design partners, people off the waitlist. And if anyone on the call is interested in joining that waitlist, you can go to alludium.ai, and there is a button in the top right hand corner to join that waitlist.

20:34 Thank you. So we've got a question from Peter. Is alludium a first mover in this third wave development? What is the biggest risk to success? We are certainly in the first mover category. We've been working on this for quite some time now. We have done deep, deep research on this space to really understand where the foundation model's position, where the intelligence there sits, and what is needed to really run agents in production. This is obviously a hot space. There is a lot of activity in this space. There's a lot of talk. But we have built a platform designed from the ground up for agents to run on and in it. We have built what we call platform agents that are able to use our platform to build agents and actually improve the platform as they go. I believe we have a very unique offering, and I believe that we are ideally positioned the right place at the right time with the right product.

21:33 Thank you, John. So Simon's asked, can you talk a bit about the marketing strategy? So I think that's one for Dave. Sure. So our go to market is specifically built around, verticalized agent packs. So we've focused on the initial area in specialized finance such as VCs, family offices, professional investors. And then the the plan is to to broaden the the agent packs from there. So we'll be looking at things like marketing, executive assistance. So it's not just the verticals, but we can look at the horizontal different teams as well. We're running rolling launch campaigns throughout the months and years, throughout '20 sorry, throughout the months across 2026, so next year. And we'll be expanding into new verticals and continuously adding new new agent packs as we go. So the plan is is very much to really nail that first beachhead and then across across the specialized finance area and then roll out throughout 2026 across that.

22:38 Thanks, Dave. Another one is, can nontechnical users really build useful agents on their own? I guess that's John. Yes. 100%. We have put a lot of work into simplifying agent building process. You would have seen this in the demo. You can create real world agents simply by chatting with our platform agents. That is a unique positioning. We have these built in specialized agents, experts that guide, configure, and validate agents for you. You describe what you want done. Our agents handle how it gets built. While there are some other platforms that are trying to bolt chat interfaces on to complex workflow processes, we have not seen any other platform that is doing this in the way we are with a platform built from the ground up to enable this. Most of the other platforms we've seen expect users to think like engineers, manually stitch things together, debug edge cases, and understand the low level nuts and bolts of things.

23:45 And this creates a huge gap between what these platforms promise and who can actually use them. We're closing that gap. We're turning agent building from a technical exercise into a guided product experience. In addition to that, for people who just want to use apps, they don't have to know anything about building. They have this dedicated chat interface, as simple as WhatsApp, available on desktop and mobile. So agents that other people have built and shared with them are available to them twenty four seven with them wherever they go. Thank you, John. So question in from Arjan. What is Iludium's USP which differentiates it from AgenTek products from big tech? So I guess that's, John again.

24:33 Really, that is that is very much the same as the previous question in a way, can people of can nontechnical people really build agents? So the process, the platform that we have built from the ground up that integrates the agents running on us with the platform itself is a real differentiator. Our platform exposes hundreds of different endpoints, different functionality. Every one of them can be used by our agents. So these agents have unprecedented access to integrate with the platform and use it to create, deploy, test, and share agents, and that is a significant USB. Thank you, John. So this one's probably for Dave. He's asked Simon's asking, is ecommerce part of the infrastructure going forward?

25:27 So we're looking at lots of different areas. As I mentioned, specialized finance is where we're looking at right now. Ecommerce is certainly on our radar. We've been looking at agents payments and experimenting in different areas in the background. We know that, you know, Stripe are looking at different areas in this already, so Spotify so Shopify. But yeah. So it's certainly something on our radar. It's certainly something that we're we're testing and looking at. It's not something we're launching in q one. That's for sure. But it's certainly something on our radar. That that's the beauty of of Iluvium. It it really can do whatever you want it to do. It's purely a case of being able to integrate it with the right platforms and then build the agent around those integrations and designing it how you want it to to to work. So if you wanted to, you know, look at how it's gonna manage your your supply chain or how it's gonna ensure that you're you've got certain stock in or you're you know, if you're buying and with buying and selling, these agents can do all that for you. And the the there really is no limit to to what these can be done can be done with these. It's just a case of where we have this on our on our road map.

26:42 Towards the end of next year, we're looking at enterprise. But, yeah, ecommerce is is certainly something that's that's on our radar. Thank you, Dave. Andrew's just asked, is any of the technology you're using in the medium patented? So it's a old chestnut in terms of technology and patents. Don't know which of you wants to have a go at that. Yeah. At the moment, we have not applied for any patents for our technology. As people may be aware, getting patents in the software space can be incredibly challenging. But by the same token, none of the underlying technology that we have built Iluteum on is patented, so we have no patent licenses that we need. We are building on the absolute best of breed open source technologies for Iluvium.

27:31 With regard to patent applications, it is something that we may look at in 2026. But our focus right now is getting this product to market, getting people on board and using us, and really showing and demonstrating the value of the product. Thank you very much. I think probably we're gonna finish in a minute on a a last question, which is what feedback have you learned from users so far? So the, the ability to have agents running in the background, these ambient agents that can be set up with triggers and work autonomously and reach out to you, has been one of the biggest wow moments, that we have received so far in terms of feedback. Another piece of feedback that we had received from one of our, design partner users was how interactive the agent configurator is. So one of our users, came in and was looking to build an agent who was doing too many things. And that that is something that we we have seen is agents perform best when they're narrowly focused. So having multiple agents, each one doing different tasks and collaborating together produces far better results than a small number of agents trying to do everything.

28:52 So in this particular scenario, the agent configurator actually pushed back on the user and suggested that what they were looking for was too broad and suggested dividing it into multiple agents and help to refine the scope of each one. And the the feedback there was that that was incredibly helpful from a nontechnical user to see the agent thinking and reasoning and coming back with those sorts of suggestions. Great. Well, thank you very much, John. Thank you, Dave, and thank you for everybody for for joining this morning. This just to remind you, this will be posted online subsequently so you can watch it and share it with other people because we want to spread the word. And, you know, we look forward to 2026 with a lot of optimism, and you can hear from what you've heard that there will be more news to come.

29:46 So thank you all very much, and see you in 2026. Cheerio. That's great. Well, thank you very much for updating investors today. Could I please ask investors not to close the session? It should now be automatically redirected to provide your feedback, and all the management team can better understand your views and expectations. On behalf of management team of CATENAI PLC, we'd like to thank you very much for attending today's presentation, and good morning to you.