Skip to content

Lab · Experiments

Experiments, in the open.

Real experiments with AI agents and multi-agent systems in business, documented openly. Each one is labelled by maturity, from a concept to something deployed, and by how much has actually been measured. Unmeasured means unmeasured: no results are claimed that weren't measured.

Each experiment maps to the AI-Native Enterprise Framework →

Stage

Concept — An idea being explored. Nothing built yet.

Prototype — Built and working in a test setting, not yet in daily use.

Pilot — Running with real work and real people, in a limited scope.

Deployed — Part of how the business runs day to day.

Evidence

Unmeasured — No results measured yet.

Anecdotal — Observations from the people involved, not measured.

Internal measurement — Measured by the team itself, not independently checked.

Verified — Measured and independently checked.

Experiments

Nothing published yet

The first write-ups are being prepared.

Experiments appear here only once they're documented honestly, with their stage and evidence. In the meantime, read the writing or see the framework.