OtonocoAI — from developer-mode prototype to bank pilots, on a pre-seed runway.

What they were building

Capital-markets compliance is one of the more painful corners of financial-services operations. Regulators publish updates constantly. Each one has to be parsed against the institution’s existing obligations, mapped to the workflows and policies that need to change, and tracked through implementation. Most of that work is still done by analysts reading documents and updating spreadsheets.

Nakhoda is OtonocoAI’s flagship product — an AI-powered platform that takes regulatory changes, analyses the obligations they create, and automates the compliance workflows that follow. The thesis is straightforward: this is work that AI is genuinely better suited to than humans, and the financial institutions that adopt it first will operate with a structural cost and accuracy advantage.

The state when I came in

The product existed but it was in developer-mode — runnable, viewable, demonstrable, but not in a state a real customer could deploy. That gap is the gap a lot of pre-seed AI startups hit. The prototype proves the model can do the thing. The next bar is whether the model can do the thing reliably, on real customer data, with the kind of guardrails an enterprise expects. Closing that gap on a pre-seed budget, with a funding decision approaching, was the constraint.

One of the founders had been in my MY2 cohort at Antler, which is part of why they reached out: they wanted product help from someone who knew what an Antler round actually demanded versus what the program told you, and who’d already navigated that round once.

What we shipped

The work covered four fronts in parallel.

The MVP definition. We reframed what the product needed to demonstrate for the round — not the full vision, but a sharp version of the workflow that worked end-to-end on customer-grade data. That focus made the round defensible.

The prototype itself. I worked on the prototype-side acceleration directly — closing the gaps between what the engineering team had shipped and what the customer demos required, so the team could keep moving without waiting on the founders for every product decision and the founders could keep moving without me blocking them.

The team. Hiring on a pre-seed budget means you can’t bring in a senior engineer at market rate. We worked through the network to identify specific calibres who could plug into specific gaps, scoped the engagements to fit the budget, and brought in the right people without overcommitting the runway. That hiring discipline is its own product decision when the runway is the binding constraint.

The customer side. We joined the pilot meetings with the early customers — Malaysia’s largest financial institution and one of the country’s top-five banks. The work there was helping the founders carry the technical conversation in the room, hear what the customer was actually asking for under the surface request, and translate the compliance requirements back into product roadmap.