I have lived compliance. Not from the outside, not as a vendor, but as someone who has sat through the loops, the back-and-forth, the lost-in-translation follow-ups, the delays, and the pressure.
The pressure comes from everywhere at once. The deal lead wants to close. The client wants to move. The compliance officer wants to do their work correctly — because that is the job, and that is what separates a decision that holds up from a decision that doesn't. Correctness takes time, and time is the one thing nobody ever seems willing to give compliance.
So what happens is a loop. The client sends partial information. Compliance asks for more. The client sends the wrong thing. Compliance asks again. 3 days pass. The deal lead is angry. The client is angry. And the compliance officer is caught in the middle — asked to move faster, but never at the expense of doing the work right.
I watched this play out deal after deal, onboarding after onboarding. We tried everything you are supposed to try. We improved our forms. We refined our risk matrix. We signed up to better screening providers. We built new communication tools. Each one was an improvement. None of them killed the loop.
That is when it clicked. You do not smooth a loop. You kill it.
That realisation became Tarth.
From aspirational idea to working agent
Tarth started out in March 2025 as an aspirational idea — half joke, half vision. An AI KYC agent that could run end-to-end client onboarding without the loops. What it is today is something else entirely. From March 2025 through August 2025, we trained Tarth against real onboardings — wiring in rulebooks, risk matrices, and policy logic by hand, cross-checking its output against the work our own compliance officers were producing, giving it feedback, adjusting it, tweaking it, polishing it, amending it, refreshing it. August 2025 was its internal launch. Since then, we have kept going. Tarth has onboarded 900+ individuals accurately. We have built out configurability so compliance teams can shape Tarth to their own rulebook, not the other way around. We have layered in multi-jurisdictional support, so a single team can run parallel regulatory contexts on a single platform. It has been through a lot.
Through all of it, we built Tarth with one rule in mind: it has to live inside the workflow a compliance officer already has. No new skill to learn. No new ceremony. No new tool to wrestle with. Tarth should feel natural the moment you use it — straight to the point, easy, built for the way the work actually gets done.
Somewhere along the way, something shifted. Tarth stopped being the thing we were building and started being the thing that was scaling us. A compliance team — small, stretched, running the same kind of work every compliance team in every regulated institution runs — could suddenly do more, without cutting corners. For the first time, we saw firsthand what a real AI KYC agent does for a compliance team when it actually works.
That was the moment Tarth deserved to become its own thing.
What we want Tarth to become
This is just the beginning. We want Tarth to serve every kind of compliance — complex or simple, high-volume or high-touch, single-jurisdiction or global. We want it to integrate with every third-party source that could enrich an onboarding decision; the more knowledge Tarth has, the better the work becomes. We want the configurability to sit in the hands of the compliance officer, not buried in a service contract. And we want to hold Tarth to the highest standard — because that is the standard compliance officers are held to, and that is the standard their clients deserve.
Tarth is for the compliance officer who wants to deliver consistently — accurately, at depth, embedded in the workflow. Tarth is for the institution that has outgrown the loop but has not yet found the way out. Tarth is for the team that wants to scale without replacing the judgment at the core of what they do.
Compliance, KYC, and AML screening are becoming more important, not less, as institutions scale and operate across more jurisdictions. The loop does not get easier with volume — it gets harder. This is the moment for a tool like Tarth: to kill the loop, embed into the workflow, and let a team scale without sacrificing judgment or depth.
We built Tarth because we lived it. We are sharing Tarth because compliance deserves better.