Microsoft's 44-page playbook: 100+ internal AI projects show licenses alone fail
alex_verem · x · 2026-09-22
Microsoft released a 44-page playbook this week distilling lessons from 100+ internal AI projects. Core finding: licensing a tool and rolling it out to 100,000 employees doesn't change how work gets done. 'A bad process with AI is still a bad process,' said transformation lead Kathleen Hogan. What worked was the opposite order: the cloud supply chain team simplified processes and built a shared data source first, then deployed 111 purpose-built agents — planning cycles that took 10 business days now take under 2.5 on measured workflows.
More from coding & agent
- Moonshot Launches Kimi Browser Extension: Record Steps Once, Agent Replays Forever — Kimi_Moonshot · 2026-09-22
- Jimothy distills LLM queries into tiny task-specific classifiers running 10-20x faster — flngr · 2026-09-22
- What Cloudflare can't do: D1 caps at 10GB, is single-threaded, and no real Postgres — Paimaamu · 2026-09-22
- User Vibecodes Tampermonkey Scripts to Patch Long-Ignored UX Bugs in University Systems — craig_macdonald · 2026-09-22
- Speaking UI layouts into existence with Jev, shadcn and a local transcription model — _AustinCalvert_ · 2026-09-22
- ai-copywriter hits 1k stars: an agent skill that writes converting copy with zero AI tells — tom_doerr · 2026-09-22