Lenny Rachitsky synthesizes interviews with leaders at six AI-forward companies — Shopify, Duolingo, Zapier, Whoop, Ramp, Intercom, and Roblox — into one blunt diagnosis: 'The biggest barrier to AI adoption isn't technology; it's organizational change.' At most companies, employees are stuck on vague AI mandates, procurement bottlenecks, and no idea which use cases to start with.
The five recurring moves AI-forward orgs make:
- Explain the how. "We are AI-first" is a slogan. Tobi Lütke's Shopify memo named specific tactics (AI prototyping in the GSD process). Wade Foster declared a code-red moment at Zapier and gave employees a week off to work through a shared playbook.
- Track and reward adoption. Shopify made AI use part of peer reviews. Ramp publishes weekly power-user counts by team. Intercom uses merged PRs as a productivity proxy. The point is to create an explicit link between using AI and what the company rewards.
- Cut the red tape. Approve multiple tools (Shopify lists Claude, Cursor, Copilot, Perplexity, Cohere, Gemini, Claude Code), let employees nominate new ones, and shorten procurement.
- Turn enthusiasts into teachers. Duolingo runs Friday two-hour AI blocks and AI Show & Tell at all-hands. Zapier hosts weekly demos that draw 60+ employees. Whoop's Hilary Gridley, when a PM brings her a problem, says: 'Want me to show you how I solve this with AI?' and shares her workflow live.
- Prioritize the high-impact tasks. 'Most companies fail at AI adoption because they try to do everything at once.' Find the high-volume, repetitive tasks consuming the most employee time and ship those first.
For a mid-level PM, the leverage is not writing the memo. It's running the demo, championing the tool, choosing the high-volume task, and being the person whose AI workflow other people copy.