As “AI-native” pitches start to fall flat with compliance software buyers, Skematic’s founders Charles and Brian sit down to talk about what actually moves the needle. They draw a line between where AI belongs, such as reporting, paper statement extraction and always-on gap analysis between your policies and your programs, and where it doesn’t, like a code of ethics rules engine that has to stay exact and auditable. They also address how Skematic stacks up against well-funded startups, why decades of hands-on compliance experience can’t be faked, and what’s next for AI-enhanced reporting.
The through-line: accountable people with better tools, built on a foundation regulators can trust.
Transcript
Charles: I spend a lot of time building product and looking at available AI solutions, and what’s available, and people’s perspective on it, is changing quickly. So, Brian, to start the conversation: what are you seeing on the front lines when people are looking to adopt technology? What are people looking for, what do they like, and what are they hesitant about at this point?
Brian: It’s evolving quickly, as you noted, Charles. When the proverbial AI bomb went off, the initial reaction was that some firms took the approach of “we absolutely need AI in the product, as much AI as possible.” Now that the fog of war is starting to clear, we’re seeing that the “AI for the sake of AI” pitch is no longer landing.
For example, one of our partners demos Skematic side by side against two AI-native platforms every day, and we were told flat out that the AI bells and whistles are no longer landing. That doesn’t mean the market isn’t looking for a thoughtful approach to AI in compliance software. But buyers are starting to realize that a lot of it is smoke and mirrors. There are use cases and functionality that can move the needle, but the approach needs to be thoughtful and product-specific rather than AI just for its own sake.
Charles: I think “AI native” is a funny concept that people don’t always interpret properly. What does AI native mean? Does it mean your system uses AI? Does it mean you built your system on AI? Those mean different things.
Brian: Right. And does that matter?
Charles: It matters for different parts of the market. If you’re building a generalized software solution for business, maybe it matters less. If you’re serving a regulated space that needs to be audit-driven, with transparency and a regulatory rules framework, it matters more.
There’s vibe coding, and there’s vibe-coded software. There’s also what we’ve seen a lot of on the market: solutions that use AI as a wrapper to solve problems without having the infrastructure in place to really understand the problem.
Brian: We’ve never tried to win on flash. That’s just not our MO at Skematic. We know the strongest thing you can give a compliance team is a program that’s auditable, defensible, and lets them pull information quickly. If AI helps with some of that, fantastic. For other pieces, it doesn’t.
Reporting, for example, is a great AI use case. You can get information out of a system of record quickly and in a customized format. Personal trading rules in a rules engine, on the other hand, are not an AI use case.
You’re the product architect, so I hope I’m not speaking out of turn, but thematically that’s a lot of what we’re hearing across the market. Sometimes it requires education. A client comes in wanting AI this and AI that, and you need to walk them through where we see AI as valid, appropriate and useful. In an area like personal trading, you can pose the question back to them: “What is AI going to help you with here?”
Charles: I think the question of can AI do this versus should AI do this, or can AI help with this, is important. I agree with you that code of ethics is highly regulated and based on practice and oversight in a defined way. We don’t want to be interpretive there. We don’t want to create inference where it isn’t needed when we can do competent analysis.
But when you look at other areas, like understanding a larger compliance footprint across firm compliance, policy management and code of ethics issue management, that’s where AI becomes really powerful.
Brian: So how is the product team incorporating AI into Skematic, and what’s the plan going forward?
Charles: The first thing we did was release the chatbot, which gives you access to your data and reporting and a current, 2026-level experience for interacting with your compliance data.
What we’re doing now is releasing more specific use cases that help compliance professionals with tasks that are labor- or time-intensive. Paper statement extraction is the perfect example of a problem AI is suited for. We’re rolling out an agent that extracts transaction and holdings data for clients, with a QA layer, and then lets them escalate that data directly to the database.
What we want to do is look at competent, available functionality and apply it thoughtfully to the compliance space. Paper statement extraction makes sense. So does deeper comparison between documents and compliance programs.
There’s the initial need when implementing a firm: extracting obligations from a document and building a calendar. It’s important that professionals work alongside the AI there, with a human in the loop, because the last thing we want is to create inference when setting up a regulated program. We want to be exact.
But that doesn’t mean every person has time, at all times, to check, reanalyze and reconfirm along the way. An AI layer that automatically understands document obligations and compares them in real time to your compliance program adds something that doesn’t exist at the human level. We don’t have time to do that every day or every week, but the AI does.
Brian: So it’s accountable people with better tools, not tools instead of people.
Charles: Right, which is how we built Skematic in the first place. We never showed up claiming to know how firms should run a better compliance program. We said, “We design tools that understand your framework. Let us give you a more compelling tool set and a framework that resonates with compliance.”
Brian, I know CCOs ask you, “What differentiates Skematic from a startup alternative that’s perceived as better funded?”
Brian: First, to dispel any concerns about Skematic’s scale and funding compared to our competitors: Skematic was acquired in 2024 by Financial Recovery Technologies, a large and very successful company that operates under Cross Country Group, a multibillion-dollar family office. In terms of funding and infrastructure, Skematic measures up, and in many ways exceeds the newer market entrants, given the maturity of the organization we’re part of. Our funding is excellent and has real depth.
Beyond that, Charles, something that has been part of our thesis for a long time, and that our clients keep validating through the customer support experience, is that you and I have been around the sun a few times. We spent a combined 20 years in this space before we started Skematic.
Many of the founders we compete with are brilliant technologists who spotted a market opportunity and went after it. We didn’t spot an opportunity. We lived in this space and understood, firsthand, the problems with the incumbent solutions for decades. So we approach this problem set from a place of perspective, with a product that reflects that perspective and, most importantly, a support and implementation team that operates the same way.
Even looking past the newer entrants to the older incumbents who’ve been here as long as we have, our support team has more experience than most of the people on their bench combined. You can’t accelerate institutional knowledge. It comes from hard-earned years in a space, and we have that in droves.
Charles: I love that, because I know from my own experience in compliance that regardless of what your demo or your AI looks like, people are looking for a partner. They want a partner who will help them move downfield. Their responsibility isn’t to flash. It’s to the regulators and their stakeholders. That’s how we’re oriented, and that’s how we continue to support clients.
We didn’t feel we had the right to run with new AI tools until we took care of our immediate responsibilities to our clients. That responsibility comes back to infrastructure and data ops, much of it around our framework and our approach to code of ethics functionality. So we’ve spent a lot of time building enterprise-level functionality that solves the task at hand.
Brian: Yes.
Charles: And what we’re really excited about now is pivoting from feature-and-function building, which set a new high-water mark, to AI enhancement that will let the tools available across sophisticated software be applied thoughtfully to the compliance space.
Brian: Absolutely. The foundation you and the team went heads-down building can support a 100-floor skyscraper.
Charles: Yeah.
Brian: Because I’ve lived this space, and the hardest question to recover from, or the one a client can’t recover from, is “we don’t have the data.” That was the vendor’s responsibility. It’s why the client partnered with a vendor, and delivering on it is our responsibility as a partner.
Some things never get old. It’s expected that your vendor knows more about your data footprint than you do. That means the vendor carries the outsize responsibility that comes with that knowledge. They shouldn’t wait for you to discover the skeletons in the closet. They should be cleaning the closet every single day.
Unfortunately, we’re seeing people in the field who may not be getting what they bargained for from their solutions, because the infrastructure, the data ops, and the understanding of that responsibility aren’t there. Vendors have a responsibility to clients, to auditors and, not to be cheeky, to the financial industry at large, and we take it very seriously.
So we use AI for extraction and for analysis, but that won’t be an interesting sentence a year from now. What’s next?
Charles: We’re going to raise the bar on reporting. Without revealing everything we have in the works, we’re using our infrastructure and footprint to elevate what people currently understand as compliance reporting, and how far the technology can take you before a human picks it up.
The other thing we’re dedicated to, as I mentioned earlier, is continual gap analysis: always-on compliance analysis between what’s codified in your policies, documents, investment agreements, vendor agreements and contracts, and what’s actually being carried out across your programs.
People only have so much time in the day. They have their to-do list, their calendar, and the things they’re doing to move forward. AI can become an overlay that understands the health of the fabric connecting all those interrelated pieces. When something comes up, it’s easy to identify, easy to resolve, and easy to incorporate back into the overall system.
