Skip to content

Innish Labs

A quiet argument
for boutique spatial AI

Most organizations don't need a platform. They need the right small team, working in the open, delivering a specific system that works. Innish Labs is a two-to-five-person practice at the seam of GIS and AI, hired by directors who've been burned by bigger firms and now want to work with people who will actually ship.

Interactive experience
Scroll to explore

The practice

Grounded in the territory

Spatial systems fail when they're designed from the org chart instead of the ground. We start every engagement in the data (parcels, networks, easements, the field notes nobody digitized) because that's where the real constraints live. A week spent with the actual records saves a quarter spent arguing about requirements.

What we do

Four ways in

Every engagement is scoped to the thing that has to work. These are the shapes the work usually takes.

Spatial retrieval

Retrieval systems that understand geography, so a zoning or asset question comes back grounded in the right place, with the source attached.

Agent operations

Agents that do real work inside real constraints: permitting queues, inspection scheduling, records requests. Boundaries first, then autonomy.

Field-to-decision systems

Getting what a crew records in the field into the hands of the person who decides, the same day, without a nightly batch job in between.

Team enablement

We'd rather make your analysts dangerous than make you dependent on us. Engagements are designed to end.

A closing note

Built to ship,
sized to care

The work we care about (the map that settles an argument, the record that survives an audit, the answer that arrives while the decision is still open) rarely comes out of a platform. It comes out of a small number of people who understand the ground, sit with the constraint, and stay until the thing runs without them. We keep the practice small on purpose.