Kanan Labs · About

About Kanan Labs

We build the intelligence infrastructure that lets firms prove their position across regulators, brokers, banks, insurers, and trade counterparties.

Regulatory volatility has become structural. Tariff regimes shift quarterly, sanctions perimeters within a week. Intelligence infrastructure is now a necessary input to compliance — not a luxury layer above it.

01 · The Service Model

We send the plumber, not the toolkit.

Most enterprise AI is sold as SaaS: a dashboard, a set of credentials, and a support manual. The customer is expected to learn it, staff it, and operate it. When the pipe bursts, the vendor ships you a better wrench. We don’t do that. When a tariff line shifts at midnight, you don’t want a new screen — you want the analyst who already knows your shipment.

The SaaS model

Sends the toolkit

Vendor Customer

You receive software. You hire the team to run it, you train them on it, and you absorb the failure when the regime moves faster than your roadmap.

The Kanan model

Sends the analyst

Customer

You receive structured operating context. The analyst is inside your workflow, the compiler is doing the labour, and the only thing crossing your desk is a decision you can act on.

Structurally, we’re closer to a senior trade-advisory firm than to a SaaS vendor — the economic model, the liability surface, the way customers buy, the trust rituals all look more like BPO and KPO than like seat licenses. Tradewatch is the first deployment. The larger undertaking is an intelligence layer for regulated trade.

02 · The Engine

Drona isn’t a wrapper on top of an LLM. It’s a brain.

Most regulatory AI tools sit one layer above a foundation model and call themselves intelligent. Drona is structurally different. The language model is one bounded subsystem inside a larger compiler — feeding a deterministic symbolic core whose outputs are traceable to source, signed, and reproducible.

Defensibility-tier output

None
Litig.
Partial
Full
Open
Non
Litig
Partial
Full
Conditional
Non
Litig
Partial
Full
Blocked
Non
Litig
Partial
Full
Unclear
Non
Litig
Partial
Full

Every output carries both axes. Doctrine inherited from the Indian Supreme Court’s reasoning in Harsolia Motors and Sohom Shipping.

Same compiled rule pack, same shipment context, same output, every time. Determinism is the architectural property the rest of the system is engineered around.

03 · The Founder

I founded Kanan Labs from a simple conviction: cross-border trade is no longer governed only by access, price, and logistics. It is increasingly governed by whether a firm can prove its position across regulators, brokers, banks, insurers, and supply-chain counterparties.

My training across multiple legal systems shaped how I understand this problem. Rules do not operate in isolation. They move through institutions, documents, incentives, and evidentiary burdens. Kanan is being built to turn that complexity into structured intelligence that firms can actually use.

My background spans law, business, and cross-jurisdictional regulatory analysis across India, the EU, the UK, the US, and Australia — with a focus on AI regulation, trade-linked legal structures, and industrial policy.

Sreenath Govindarajan

Founder, Kanan Labs

Cross-border law · AI governance · Regulatory strategy · Trade intelligence


What I believe

  1. Architectural seriousness as posture. Quick wins that compromise the architecture are rejected.

  2. Defensibility, deterministic-first evaluation, and continuous compounding are non-negotiable.

  3. Direct challenge welcomed over passive agreement.

  4. Low tolerance for conceptual conflation.

  5. File-based memory as constitutional substrate. The company’s memory is on disk — portable, auditable, gittable.