Untitled - June 4, 2026

00:00:00 Speaker: Hi, everyone. This is Ed Mcquiston, CEO of KeyMark. And this is the Mostly Unstructured Podcast. And I am here with Matisha Ladiwala from Hyland who is running all of the product side for CIC. And, uh, first of all, thank you for joining the podcast. Um, so I think you told me you've been at Hyland six months. So how's it going? Yeah, thanks for having me. Happy to be here. Um, I joined Hyland six months ago from Microsoft, and, uh, I have not regretted it one day. That's excellent. Super excited about what we're doing here, the opportunity that we have ahead of us. You know, I think just the time is right for what we're doing here at Hyland. You know, the technology exists to enable us to do things with unstructured data that, you know, quite frankly, we couldn't even dream of, you know, two years ago. Right. So, you know, I continue to be very excited with the opportunity. And it's just been and been an amazing six months so far. And, you know, as you've come in these first six months, you've had a chance to kind of see under the covers, you know, what do we have? What are we doing? What's what's excited you the most? So, you know, it's interesting, um, you know, today it's, you know, anybody can build an agent, right? Everyone is actually building an agent. So I think, you know, very quickly, it's not going to be about, you know, who can build the fastest agent or who can, you know, tell the best story even. I think it's about it's going to go back to fundamentals. It's going to be about data, you know, content. It's going to be about, you know, you know, security, trust governance, you know, those kind of basic foundational elements. That's really what's going to become important. You mentioned governance. And, um, I had a meeting with Mike Campbell, Chief Product Officer, a couple months ago, and I asked him somewhat similar to what you're saying. Like, what's what's different about CIC versus the person who's going to say, well, I can vibe code, you know, an agent to go get unstructured data. And he said, you know, if I had to distill it in a word, governance. So you want to talk a little bit about why he was, you know, so quick with that answer. Yeah for sure. I mean, if you think about the history of Hyland as you know, you know, Hyland has built its business on, um, on industries that need, uh, governance, their data that need to, you know, trust that the vendor that they're, you know, entrusting with their data is going to help them be compliant, secure. And we've you know, if you ask, you know, what has Hyland done for the last thirty five years? Like we've really built trust exactly with our with our customers. And so, you know, uh, as we think about governance, as we think about AI, you know, extending that trust to, you know, now the AI agents that you build, right? You know, you know, one side of it is, you know, using the content to power the agent. But the other side of it is, you know, trusting that the agent is actually it's not going to hallucinate. It's not going to, you know, do things with your data that you don't want it to do. Right. Um, and so it's really an extension of our initial value prop to AI now. Um, and yeah, I would agree, I would say the thing that, that, you know, makes AI at Hyland different is, is that trust and governance piece. I totally agree. I mean, I had a meeting with our, our customers yesterday and I talked about all the time from the moment you demonstrate to a customer that you spend talking about the security and governance of the unstructured content, of who gets to see what in what scenarios. And they can see these documents all the time, except for these occasions and the time and effort that gets built into that. And so, you know, when you contemplate, you know, oh, I'll just use some third party agent, you've stripped away what you've put years of effort into making sure is secure. And Hyland of course, works. And as KeyMark does, heavily regulated industries, healthcare, financial services, you know, insurance, government, you you just can't say, oh, well, you know, we'll just let some third party come in and, and so I think, you know, that that that answer resonated with me a lot because, you know, the amount of time and effort that people put into those security models, it matters. Mhm. So, you know, as you're, you know, designing, you know, where, where you're going next as you look out at the next, I don't know, these days, I feel like the, the time horizon gets shorter and shorter because everything's moving so fast. If you look out over the next six to twelve months, what what are you seeing? What are you excited about? What's kind of the where's your North Star at the moment for CIC? I think, you know, just looking at where customers are, I think this week here at CommunityLIVE, you know I've had an opportunity to talk to many customers, you know, get their reactions to some of the things that we're putting out on our roadmap, you know, with kind of, you know, the context and the governance and what have you. And what I'm still seeing is that there's a large number of customers that are still in that pilot phase with AI. Yeah. And so I think what's going to be critical over the next few months is how do we help customers, you know, build trust and, you know, take some of these pilot projects into production. Yes. You know, and start the, you know, actually see more and more ROI. Uh, but you can't do that unless they actually, they actually kind of trust the output of your, of, of your AI agents. And, you know, like if you get to the bottom of it, you know, what is trust? Trust is just the ability to have predictable outcomes, right? So the more you use the agent and the more you see that it is delivering outcomes that you can trust that are predictable, the more people are going to use this. And so that's really what I would like the focus to be for the next few months is let's get folks starting to, you know, really double down on the usage of AI and start the, you know, actually realize some of these gains that we've been talking about for the last few months to your exact point. So I said in a bit on the executive summit here at CommunityLIVE and, uh, you know, Jitesh Hyland CEO was interviewing, uh, Shankar from Great Ormond Street Hospital who runs, I think he's, I don't know the exact title, but runs AI for the whole of of NHS. And he sort of that was his plea to the room stop piloting. Yeah right. Put it, put it through its paces like you would any other project. I mean, I think that's probably the big difference is there's, you know, there's a willingness to try technology in so many sectors. You know, maybe it's a short pilot and then go live. And I think there has been a bit of that, you know, hesitancy. And, you know, you talked about, you know, I'd love to actually kind of go back to this idea of the technologies that we have today when it comes to unstructured that just didn't exist before. And, you know, I'm, I've been around this industry too long. I had hair when I started. And one of the things that amazes me right now is the technology that exists that you used to just not be able to contemplate. You go, all right, well, this whole setup, there's no way we can read that. There's no way we can deal with it. Like just keeping up now with the technology pace. And how do you how do you do it? Yeah, I mean, it's, it's a full time job, isn't it? Isn't it? Um, you know, I think we're very fortunate at Hyland you know, other than kind of keeping myself up to date with kind of what's going on, reading publications, talking to analysts, looking at what other companies are doing. You know, I think we're fortunate at Hyland that we have a very, uh, you know, we have kind of a whole team that's focused on, on AI, right? Uh, you know, the thought leadership aspect of it, you know, experimenting with, uh, you know, with new things continuously. So we're always kind of a, you know, we try to be one step ahead of, you know, of, of, of other competitors that are out there. But, you know, honestly, right now, you know, even as we're building this, a lot of it is, you know, is around experimentation also. Sure. Um, and when I say experimentation, I mean, it's, you know, Almost. I find that the most successful things we put into market are ones that we're actually closely partnering with, you know, a customer and jointly kind of, you know, building to a specific value prop, you know, showcasing, you know, ROI immediately and then kind of scaling that out to a larger group of group of customers. But, you know, I think to your point, like speed, both of, you know, speed of learning, speed of change, uh, speed of development, speed of deployment is just critical. Yeah. Because things are changing so quickly. One of the, one of the things that I know for me, you know, I last, um, you know, I joined KeyMark just over a year ago. And, um, so we did a big kickoff, you know, kind of all employee type thing. And, you know, I talked about sort of like, hey, this is, this is how I am seeing the market. One of the things that I sort of in that moment, this is literally July, just a year ago, less than a year ago, I was looking at the data lake side and I said, okay, I think they become the one ring to rule them all because that's where the data is probably going to end up. And so that's where people are going to be building agents and da da da da. And then I read about MCP, which I don't even know if that's a year plus old at this point. And it blew up my whole perception of the market because now those Hyland agents can communicate to any number of endpoints to include as a semantic layer to enhance. Like, is that something that you guys are talking about? Absolutely. You know, and I think the reason it becomes even more important is, you know, I think it would be, you know, vein for, for us as Hyland to assume that we have, you know, that that, that all the company's data or all a company's agents are going to exist on our platform. It's just not true. Right. And so, you know, two core pieces of our strategy. You know, if I think about the data layer, uh, is, is first federation. So, you know, federating the data from third party sources, as we call it, you know, systems of record with the content that is stored in Hyland right? You know, that's kind of the step one. And then the second concept that we talk about is this concept of the agent mesh, which is really a group of agents working together to accomplish a specific business outcome. And again, not all those agents are going to necessarily be ones that you build on the Hyland platform. And so that's where things like MCP come in, where it's not just, you know, now kind of interactivity of data, but it's interactivity of agents that becomes critical in order to really, you know, deliver that business value. Absolutely. So it's definitely a focus area for us. Um, and, you know, I think as, as, as time goes on, you'll see more kind of use cases that we put out that, you know, leverage, you know, this combination of agents within Hyland and, you know, really anywhere that they're built. One more thing that I want to say on that topic is, you know, one of the things that we announced this week is, um, kind of headless access to our, uh, data and context layers. And so what this means is now even if you build the agent kind of in a, you know, in a, in a different system, right? Uh, you could still kind of leverage the data, you know, the ontologies, the context that we have in Hyland to ultimately, you know, power that agent, right? Right, right. It's going to become not, you know, how many agents do I have? But it's which agents are giving me the best outcome. Yes. And those are going to be agents that leverage data across systems, agents that live in multiple systems, but are all kind of focused on achieving a specific business outcome. So you mentioned something that caught my ear and, you know, maybe, maybe two more questions that, um, you talked about. Federation. Yes. And I know when, when I first saw the kind of CIC vision, I got super excited about that because to your point, you know, having been in Ireland a long time, as much as we'd like to think that we're the center of the universe a little bit, there's a lot of ecosystem there. And it's one of the things that really struck me is I think that from the federation standpoint, it's, you know, it's more than just other unstructured silos, right? It is this idea of both unstructured but also discrete data from different endpoints. And that that got me really excited because I think that's, that's the piece. And, you know, I remember talking to Jitesh at the time and he talked about, you know, where he came from, which was more the discrete data world. He looked at this, this was sort of the final frontier. This was the last mile. And that federation really allows you to to tie it all together. I don't know if you have any comment on that. I couldn't agree more. And I think, you know, more and more in this, uh, in this time that we're in now where we're looking for measurable outcomes from, you know, data and AI, it is, you know, and as we look about, you know, look at automating workflows even, yeah, you know, we're parts of the workflow lives in one system. It's fueled by data that lives in a different system. You know, absolutely. It's going to be, and it is about bringing together, I'll call it, structured data from systems of record and unstructured data, you know, to really kind of fuel that next level of automation experience output, you know, business results, whatever you want to call it. Uh, but I think, you know, you know, this is when I say federation, it's both kind of the structured and this unstructured data that's really where the power of this is. And maybe one last question. So one of the things that was very noticeable to me yesterday, which I think is a great, uh, evolution in your strategy is. Hyland has been so successful over the years because of its vertical implementations, right? The deep domain expertise in healthcare and government and financial services, etc.. And one of the things that I think can be a huge differentiator for Hyland is most of what what's out there is generic platform level AI, right? So great technology, great agents, nothing wrong with it, but no kind of built in domain knowledge. And what I saw yesterday appears to be the front end of a strategy to say, we're going to pre-build some smarts because we know the processes, we know the customers, and we're going to kind of create some, you know, pseudo off the shelf type offerings. Can you comment on that? Right, right. So it all starts at the data layer, right. So you know, yes, you know, we've, we've, we've kind of worked with, you know, some of these industries for, you know, many years, thirty five years. Uh, we understand what their data structures look like, right? And so the first thing that this powers is, uh, you know, what we've been calling the enterprise context engine, which is really industry specific knowledge graphs, is what it creates. It enables us to create. And now this is what powers kind of agentic solutions. And so, you know, yes, it's not just about, you know, hey, we have a solution for healthcare. We have a solution for financial services. But it's a, you know, we're bringing kind of this, this, this, you know, domain specific expertise to the knowledge graph first and then to the specific process. So, you know, it's no longer about, uh, you know, you know, where do I start? How do I get value out of this? Exactly. I think Jitesh used the term yesterday at the keynote, which is, you know, the era of ubiquitous noise. Yes. And so this is our attempt to say, you know, we're gonna we're gonna try and simplify all these layers for you. You know, you have the data in your, you know, in your content management system. We're going to give you out of the box what this ontology should look like, which of course will improve over time as you use it and you provide feedback. And we're going to give you an out-of-the-box solution to, you know, a process that we know that, you know, has an opportunity for, uh, you know, for, for improvement or automation so that, you know, you're not struggling for months or even weeks to, to start to get value, but you can start to get gradual value. Um, and what this also does is it helps with this whole AI adoption journey. Absolutely. Right. Start with a small process, um, you know, expand it to more processes, but it kind of also, you know, coming back to kind of the kind of one of the key points that I've been trying to land in this is this concept of trust, right? You know, if you can use it for a specific use case that's specific to your industry, you're getting good results out of it. You're going to build trust. And you know, your AI adoption as a whole is going to, you know, increase in your organization. I completely agree. I mean, I think you accomplish you get speed to value. There should be a willingness to adopt and move out of pilot much more quickly, because it's clearly built for this process. And I think maybe as or even more importantly, you take away the customer having to sort of assimilate this generic or horizontal technology to their business problem. You go, no, no, here's this is how we're solving it so well. You've been so gracious with your time. We genuinely appreciate you being on, um, you know, I want to thank and thank you for joining the Mostly Unstructured Podcast and we'll see you next time. Thanks.