Strategy

Where small businesses should actually start with AI

Most businesses we meet have the same two feelings about AI at once: a nagging sense they should be doing something, and no idea where to begin. Both are reasonable. Here's how we cut through it.

Start with the work, not the technology

The mistake is starting from "what can AI do?" The better question is "where does my team lose the most time to repetitive, rules-based work?" That's where AI pays off first — and it's almost always document- and communication-heavy work: drafting, reviewing, summarizing, responding, routing.

If a task is repetitive, language-heavy, and follows patterns your team already knows, it's a candidate.

Rank by value, then by effort

List the candidates. For each, estimate two things: how much time it eats per week, and how painful it would be to get wrong. The sweet spot for a first project is high-time-savings, moderate-stakes — meaningful enough to matter, forgiving enough to learn on.

Keep a human in the loop

The goal of a first project isn't to remove people — it's to remove drudgery. The best early systems do the heavy lifting and hand a human the final call. That builds trust, and trust is what makes the second and third projects easy.

Ship something small that works

One working workflow beats a grand strategy deck every time. A small, real win creates momentum, surfaces the next opportunity, and teaches your team what good looks like. That's the whole game in the first 90 days.

---

← All insights