Morrisons AI Solutions: intelligence today, a brighter tomorrow

Product · not for sale yet

Build Intelligence

A service that helps an AI agent decide what to build and test, run it safely, and tell the difference between “the tests passed” and “this is actually verified”.

Design stage

What it is

When an AI agent changes code, a green test run is often taken as the end of the story. Build Intelligence is designed to ask the harder questions: what should be built and tested, did the right things actually run, and is this change really verified?

Plans and runs

It works out what to build and test for a change, then runs it in a controlled way.

Explains failures

It sorts failures into the real ones, the flaky ones and the ones caused by the environment, and says which to fix first.

Keeps the evidence

Each verdict comes with the evidence behind it, kept so it can be inspected later.

How it is designed to work

Six tools, and deliberately no tool that runs arbitrary shell commands.

ToolWhat it does
build.openOpens a build context for a project and a snapshot of its source.
build.planPlans what to build and test, and why.
build.runRuns the plan in a controlled way.
build.diagnoseExplains what failed, groups related failures, and ranks them.
build.verifyDecides whether the change is verified, and says whether the evidence is current.
build.evidenceReturns the evidence behind any verdict.

Why it works this way

Where it is today

Design only. This exists as a written design and a fixed first-version contract. There is no working software yet, and we are not giving a release date or a price. We will say so on this page when that changes.

Want to hear when it moves on? Write to morrisonsaisolutions@gmail.com.

← All our products