01 · What a strategy isn't.
Walk into any 12-person business that's just paid a consultancy for an "AI strategy" and you'll find the same artifact. Eighty-three slides, two of which the owner has actually read. A maturity model. A capability map. Six recommended initiatives, none of which fit the actual operating cadence of a business with no dedicated engineering function.
Almost none of it survives contact with Monday morning.
The fundamental error is treating SMB AI strategy as a smaller version of enterprise AI strategy. It isn't. The constraints are different, the people involved are different, and most importantly, the decision-making rhythm is different. A regional dental group doesn't need an "AI center of excellence." It needs to know which three things to try first.
The version that ships in 30 days is shorter, sharper, and assumes someone — namely you — will actually use it.
02 · The 30-day version.
The version we ship looks more like a one-page surgical plan than a strategy document. Four weeks of work, four artifacts, no extra theatre.
- Week 1 — Discovery. We interview the owner and a few key staff. We map the actual workflows, not the org-chart ideal of them.
- Week 2 — Inventory. We look at the tools you already pay for. We find the friction points and the dead air between systems.
- Week 3 — Opportunities. We rank candidate AI moves by payback, effort, and risk. Most are small. A few are bigger.
- Week 4 — Roadmap. One page. Quarter by quarter. Something you can hand to anyone — including us, if you want to do the building too.
03 · Anatomy of a roadmap.
A roadmap that survives Monday morning has three properties: it's short, it's owned, and it's sequenced. Short means it fits on one page. Owned means there's a name next to each item — and that name isn't the consultancy. Sequenced means we've made a real call about what to do first.
Acme Auto Repair · Tucson · Quarterly plan
| Q3 | Inbound voicemail → CRM ticket | Highest payback · Low effort |
| Q4 | Photo-to-quote estimator | High payback · Medium effort |
| Q1 | Auto review-ask post-service | High payback · Low effort |
| Q2 | Inventory reorder agent | Medium payback · Medium effort |
04 · What we cut.
Everything the big consultancies put in the deck and the SMB owner doesn't need. No maturity model. No capability map. No vendor evaluation matrix for tools you'll never buy. No "AI center of excellence" recommendation for a business with twelve employees.
What gets kept is the part where someone made a decision about what to try first — and then stuck around long enough to help build it.
05 · How to actually use it.
Print the roadmap. Tape it next to your monitor. The next time someone tries to sell you "transformation," check what they're proposing against the one page on your wall. If it isn't already on there, you probably don't need it.