An AI playbook for small business owners
By Dobbin Bookman
Dobbin Bookman is the Director of AI Initiatives at ICIC, where he leads efforts to harness the transformative potential of artificial intelligence to drive inclusive economic prosperity for small businesses and under-resourced communities.
We are seeing more and more small business owners integrating artificial intelligence into their daily business operations. In fact, Small Business Majority recently found that only one-fifth of entrepreneurs aren’t using AI, compared to 36% in the first quarter of 2026. While many people still have concerns about AI tools, it is becoming unavoidable.
At ICIC, our research into AI adoption among small businesses, combined with our hands-on work with business owners across the country, has given us valuable insights into both the opportunities AI presents and the challenges of putting it to work effectively, and, after working with hundreds of small business owners across the country, I want to share with you what I have learned about how to use AI to help you run your business.
How AI goes from a novelty to functional tool
Think about where you were two years ago with AI. Most of us were in the “wow” phase, copying and pasting things into ChatGPT just to see what it could do. Maybe you drafted an email or asked it to create an image. Fun? Sure. Operational? Not so much.
Then, sometime between 2024 and 2025, something shifted. AI went from a party trick to the potential to be an actual tool. You may have started delegating real tasks, drafting RFPs, analyzing customer feedback or building job posts. You started to see real use cases.
Now it’s 2026, and the operators who’ve pulled ahead aren’t just prompting better. They’re building systems. They’re turning one-off AI outputs into repeatable templates, then into standard operating procedures (SOP), then into review loops that make their entire operation sharper. These operators are experiencing increased efficiency and overall workflow optimization that ultimately leads to higher performance.
That’s what the 12 Moves you will discover below are about.
The 12 Power Moves: A framework for AI
I developed these 12 Moves through my own experience with AI, and from workshops, client sessions and live facilitation with real business owners.
“AI won’t replace strategic operators. But operators who use AI will outperform those who don’t.”
The Moves are organized into three groups, and each group represents a shift in how you relate to AI:
Group 1: Precision — Make AI know what job it’s doing
The most common mistake I see? People open ChatGPT and type something vague like “help me write a marketing plan for my HVAC business.” That’s not an assignment, that’s a wish. Precision is about giving AI enough to actually perform.
Move 1: Name the job. Don’t ask for help, assign a role. Instead of “give me marketing ideas,” try: “Act as a growth marketer for a 20-person HVAC company in the Midwest.” When you name the job, AI stops guessing and starts working.
Move 2: Define “done.” What does a good output look like? Tell it. “Create 3 options. Include cost, effort, risk, and the first 3 steps for each.” If you don’t define the finish line, AI will write you a novel when you needed a memo.
Move 3: Add context. Include industry, company size, constraints, audience, geography, and your website URL. Context is the difference between a generic answer and a useful one. When I add context to a prompt, the output quality improves dramatically, every time.
Move 4: Show your inputs. This is the unlock. Stop requiring AI to guess your business. Paste in your actual customer complaint logs, staff schedules, meeting notes, financial snapshots, job descriptions, inventory data and whatever’s relevant. AI gets 10x better when it has your real information to work with, not hypotheticals. This single move transforms output from “interesting” to “actionable.”
Group 2: Thinking — Make AI think like an operator
Once AI knows the job, the next step is getting it to think the way you think, as an owner/operator weighing options, surfacing risks, and making decisions with tradeoffs in mind.
Move 5: Constrain the format. “Return as a table.” “Keep it under 200 words.” “Give me bullets, not paragraphs.” Constraints are clarity. You need useful information, not more information.
Move 6: Surface assumptions. This one is underrated. Add the line: “List the assumptions you made.” Even better, ask AI to act as your critical mentor: “Push back on my thinking, surface assumptions, flag risks, and tell me what I may be missing.” You’ll be surprised how often AI is operating on assumptions you’d never accept, or politely validating a direction that needs more scrutiny. This move helps prevent bad strategic decisions before they happen.
Move 7: Ask for options, not answers. Never ask AI “What should I do?” Always ask: “Give me 3 approaches — conservative, balanced, and aggressive. Show tradeoffs for each.” You’re the decision-maker. AI is the analyst. Keep it that way (for now).
Move 8: Force tradeoffs. For every option, require: cost, risk, difficulty, and ROI timeline. This is the move that turns a response from Claude or ChatGPT into something you’d actually present to your team or your board.
Group 3: Operationalizing — Make AI part of your operating system
This is where the real leverage lives. Groups one and two will get you a great output. Group three turns that output into a system.
Move 9: Improve the first draft. Never stop at version one. Say: “Make it 30% shorter.” “Show me three more drafts.” “Now rewrite this for a non-technical audience.” Iteration is where good becomes great.
Move 10: Turn output into template. If you use a prompt twice, template it. That follow-up email that worked? That job post that attracted great candidates? Save the prompt structure. Don’t “threepeat” work.
Move 11: Turn template into SOP. Document who runs it, when they run it, what inputs are needed, and what good output looks like. Replicate it and propagate it. Don’t duplicate it.
Move 12: Build a review loop. Every AI output needs a human checkpoint. Is the price correct? Does it sound like us? Does it make a promise we can keep? Check for accuracy, bias, compliance, and tone.
The 12 Moves in action: Two scenarios
These are the kinds of problems real small business owners bring to my workshops. Notice how each scenario layers multiple Moves together.
Scenario 1: 40 customer complaints → 60-day action plan
The old way: Three hours reading through complaints, guessing at patterns and drafting a plan from scratch.
The new way: 10 minutes.
“Act as an operations analyst for a quick-service restaurant with two locations in North Carolina. Goal: Reduce complaints 30% in 60 days. Here are 40 complaints: [paste actual files]. Categorize themes. Identify root causes. Recommend three fixes with effort, cost, and risk. Output as a 60-day action plan with weekly milestones.”
That’s Moves 1, 2, 3, 4, 5, 7, and 8 all working together in one prompt.
Scenario 2: From job post to 90-day onboarding plan in one prompt
You need to hire. You don’t have an HR department.
“Act as an HR leader for a 25-person landscaping business. Create a complete hiring package for a crew lead: Job post (less than 250 words), interview guide (10 questions), scoring rubric (1–5 scale), 30-60-90 day onboarding plan. Context: four crews, role manages 5–6 people, 3+ years required, bilingual preferred, clean driving record. Paying up to $30/hour. Here’s a link to our website.”
One prompt. Four deliverables. Rinse and repeat for every role you hire.
The real pro tip: One task at a time
Forget “implementing AI” or building an “AI strategy.” You don’t have a strategy for your hammer or your calculator. AI is just a tool for getting real work done faster.
Instead, do this:
- Pick one task you repeat weekly.
- Build a prompt using just two of the 12 Moves.
- Test the output on real work.
- If it works, template it. Then rinse and repeat.
My recommendation? Start with Move 4 and Move 7. Those two moves alone will change your relationship with AI overnight.
Check out ICIC’s Harnessing AI portal, and learn more about AI prompting in their blog, Talk With It: The Pro Tip to Smarter AI Output.
Let us know what your pain points are around AI and what kind of content you’d like to see more of.