A canal with six lock gates in a row, each worked by its own keeper in silhouette, where a barge is cleared of floating debris before it enters a motorised boat lift at the far end, the whole scene drawn with wide empty margins on every side.

Microsoft Inside Track

Deploying AI with employee councils: An IT playbook from Microsoft

Jason Kellington · 29 October 2026

Company story · decisions · rituals · Evals before the build

Takeaway

Microsoft's internal IT organisation replaced scattered AI experiments with six employee councils that choose projects, fix processes before automating them, review responsible use and measure results.

Summary

Microsoft Digital, the company's internal IT organisation, describes how its AI work moved from broad experiments to a connected set of six employee councils. The early experiments created excitement but also duplicated work, uneven governance and gaps in reporting. The playbook, written by Jason Kellington, sets out what each council does and what Microsoft Digital learned.

Key points

A continuous improvement council asks teams to find where work gets stuck and where effort is wasted before they automate a process with AI.
A strategy council chooses the most valuable AI projects and their measures of success, while product management teams keep ownership of delivery.
Teams register each AI project and complete a responsible-use assessment before work starts, with trained colleagues acting as reviewers and advisers.
A measurement council asks for a baseline, telemetry and agreed measures as part of the investment decision, and asks where saved hours were reinvested.

Implication

A product organisation can agree how an AI project will be measured, and record the workflow as it is, before the project is funded. This is Microsoft's account of its own practice, and Microsoft sells the AI tools it uses.

Plays

Product

Before a team automates a workflow with AI, list each step as it runs today, mark the steps where work gets stuck or effort is wasted, and fix those steps first.

Product

Record how long a workflow takes and how many people it needs before an AI project starts, so the result can be compared against that baseline.

Derived by Working Surface from the article; more in the Playbook. Source line: Teams need to understand the processes affected, including how long work takes, how many resources are involved, and what the workflow looks like before and after AI.

Source issue

30 October 2026: as AI agents do more of the building, teams decide on purpose which work people keep and set up tests, usage figures and baselines before handing over more.