Generic "AI rollout best practices" miss that the failure modes change dramatically by company size. A 15-person company fails differently than a 150-person one. Here's the size-specific guide.
At <20 people: too much DIY, no documented learnings. At 20-50: too many tools, no named owner. At 50-150: committee paralysis, plans that never ship. At 150-500: enterprise paralysis, IT bottlenecks, vendor-driven instead of business-driven. Diagnose by size; fix specifically.
At 500+, the playbook is genuinely different. Different governance, different tooling, different risk profile, different vendor relationships. We are not the right advisors for this segment; engagements at this scale should go to firms specialized in enterprise rollout.
50-150 person range. Too big for founder-led, too small for enterprise structure. Committee paralysis is the most common failure mode.
No. Failure modes scale with size. Borrowing a 500-person playbook at 50 people produces paralysis.
Plan for the stage you'll be in 12 months from now, not today.
Almost never at <100 people. Usually right at 250+. The window in between is judgment-call.
Yes — sometimes the most valuable thing a consultant does is steer you away from over- or under-scoped rollouts.