Skip to content
Start a review

Make better AI bets.

Jidoka Labs is an applied AI research and operating lab. We help leaders identify useful opportunities, avoid costly distractions, and carry the right AI work into operation.

Request an AI decision review See how we work

ResearchAdviseBuildOperate

Starting AI work is easy. Choosing what should survive is not.

Every portfolio company has AI proposals. Few have a clear way to decide which deserve capital.

A weak decision costs time, capital, and management attention. A clear reason to build, test, or wait is a useful result.

Start with one written decision.

We examine one important AI question with your leadership team. You leave with a record of the evidence, uncertainty, action, cost, and operating owner.

Build

The evidence supports direct delivery.

Test

A focused experiment can resolve the key uncertainty.

Wait

The timing, evidence, or operating fit is not ready.

Request an AI decision review

One loop, from research to operation.

Enter where the problem starts. Continue only as far as the evidence supports.

01

Research

Track important changes. Separate evidence from assumption.

Output: a position
02

Advise

Apply the research to the business. Define the decision.

Output: a decision
03

Build

Work close to the team and users. Build the smallest system that can test the decision.

Output: a working system
04

Operate

Own, monitor, and improve the system. Continue, change, or stop.

Output: an operating result

Useful AI work obeys a few hard rules.

  1. 01

    A clear decision to wait can be worth as much as a decision to build.

  2. 02

    Working software is a strong test of strategy.

  3. 03

    Results from the field must return to the research.

Three ways to start.

Each path uses the same research-to-operation model.

A / Portfolio

Portfolio AI and technology review

Assess AI and technical work across a portfolio company. Set priorities and deliver selected improvements.

Discuss a portfolio company
B / Product

Forward-deployed product engineering

Build AI products, internal systems, and integrations with senior engineers embedded in the work.

Discuss a product
C / Operation

Embedded CTO and managed engineering

Add technical leadership and ongoing engineering during a startup or company transition.

Discuss an operating need

For leaders who want to improve their own work with AI: private executive AI sessions.

Jido is where the method meets code.

Reliable AI systems need supervision, recovery, and human control. Jido is our open-source framework for these systems, built on Elixir and the BEAM.

Jido makes our thinking public and testable.

Explore Jido on GitHub

Research is useful when it changes a decision.

We turn changes in AI into positions that leaders can examine and use. Field results return to the research.

Position paper

AI in operating companies

Where leaders should act, test, or wait.

Forthcoming
Field notes

From research to operation

Short reports on decisions, systems, and operating constraints.

Forthcoming

Bring us one AI decision that matters.

Tell us what you are considering and what makes the decision difficult. We will tell you whether an AI decision review is the right first step.