You don’t need a 12-month strategy to get started. You need 90 days of disciplined motion. Here’s the exact plan we walk through with new AI Office clients on day one.
Days 1–30: Foundation + first build
Week 1: Kickoff. Map your systems, data, top three operational frustrations, and top three ambitions. Identify the Sponsor (CEO/COO) and the Champion (director-level, daily ownership). If you can’t name the Champion, do that work first.
Week 2: 1-page roadmap. 5–7 prioritized opportunities scored on impact, feasibility, and speed-to-value. Pick the smallest, fastest-payback option as your 90-day priority. Don’t pick the most ambitious — that’s a year-2 problem.
Week 3: First working session on the 90-day priority. Scope the build in detail. The engineer joins the room. Start shipping rough.
Week 4: First working artifact in a production-adjacent state, even if it’s narrow. Document the baseline KPI you’re trying to move.
Days 31–60: Ship + measure
Weeks 5–6: Refine the build based on first contact with real data. Catch the edge cases. Update the workflow design.
Weeks 7–8: Production cutover. A working solution in your environment, your team using it on real work, monitoring in place, KPI being measured.
Critical: document ROI as you go, in your CFO’s language. Not “the AI is working” — actual hours saved, actual cycle time cut, actual deals won faster. Numbers that show up in the financials.
Days 61–90: Adopt + plan next
Weeks 9–10: Adoption work. Train every person who’ll touch the workflow. Get a feedback loop running, with corrections fed back to improve the next iteration.
Weeks 11–12: Review against baseline. KPI verified or course-corrected. Internal communication: this is what we shipped, this is what it produced, this is what’s next.
Days 88–90: Pick the next priority from the roadmap. Run the cycle again.
What this looks like with an AI Office partner
Most owners don’t run this alone. The partner relationship absorbs the parts that need senior judgment — what to prioritize, how to scope, what to measure, how to communicate the change.
- Sherpa ($2,500/mo) is the lightest version: a senior strategist weekly, and you drive the build internally (or scope it as a Value Sprint).
- Operator ($5,000/mo) adds dedicated engineering capacity — you don’t have to recruit a builder.
- Embedded ($10,000/mo) is for owners who want the active build queue plus exec coaching.
Total 90-day spend (Operator tier + a small Value Sprint): $15K retainer + a Value Sprint (most run $2K–$25K) lands well under $40K — a fraction of what a national consultancy engagement or a full-time senior hire would cost over the same period.
What changes after the first 90 days
Three things show up in months 4–6:
-
The team trusts the process. Skeptics become advocates because they saw something ship. That’s worth more than any sales pitch you could’ve given.
-
The next opportunity is obvious. Running the cycle once reveals the next workflow worth attacking. You stop guessing.
-
The pattern compounds. The second 90-day cycle ships faster than the first. By cycle three, you have a real AI capability in the business.
What doesn’t change
What doesn’t change in 90 days:
- Your culture. Culture moves slower than workflows.
- Your tech stack. You don’t rip and replace. You operate on top of what you have.
- Your business model. AI doesn’t fix a broken business model. If something deeper is wrong, fix that first.
If you’re hoping AI will solve a strategic problem you’ve been avoiding, it won’t. AI is leverage on an operating model that already works. Make sure the operating model works first.
When 30/60/90 is the wrong frame
Two situations where this plan doesn’t apply:
You’re not ready for a 90-day commitment. Start with one of our free or low-cost on-ramps, or AI Office Sherpa ($2,500/mo) — a lighter touch designed to surface the right starting point without committing to a fuller retainer.
Your bottleneck requires a data-readiness Value Sprint before any AI build. Sometimes the first 90 days is “get the data clean enough for AI to work.” That’s fine. It’s a real Value Sprint, not a delay. The AI work comes in the second 90 days.
Where to start
If you can name the workflow you’d start with on day 1 and want a second opinion on the 90-day plan for your specific situation, that’s the conversation to have.