Is Your Small Business Ready for AI? A 20-Question Checklist
November 7, 2026 · 9 min read · By Brad Strawbridge
What "ready for AI" means for a small business
Being ready for AI does not mean having a data team or a large budget. It means you can answer four plain questions: What problem are we solving? Can we find the information it needs? Who owns the result? What rules keep it from making a costly mistake? AWS describes an AI readiness assessment as a structured look at your data, technology, processes, and people, and says its purpose is to prevent wasted spend and show gaps early. We add two things owners often skip: written rules, and a way to prove the project is working. A business that is not ready is not doomed. It just has a short list of things to fix before buying anything.
Sources: AWS, AI readiness checklist for SMBs (September 1, 2025)
How to use this checklist
There are 20 yes or no questions in five areas, four questions each. Answer yes only if you can show it today, such as a document, a report, or a named person. "We plan to" counts as no. Give yourself one point for every yes, then use the scoring table below. Ideally the owner, the person who runs operations, and someone who does the work all answer together, because they will disagree about a few of these, and the disagreement is useful.
Area 1: One clear business problem
Start with the problem, not the tool. AWS advises defining one business outcome for your first pilot and how you will measure it, and choosing work that is repetitive, high volume, and clearly defined. For a service or home-services business, that is usually something like slow replies to new leads, quotes that sit unsent, or invoices that go unchased.
- Can you name one repetitive task or leak, such as missed calls, slow quotes, late invoices, or unanswered follow-ups, that costs you time or money every week?
- Do you know roughly how often it happens and how long it takes today?
- Can you say in one sentence what better looks like, for example "reply to every new lead within 10 minutes"?
- Would the result show up in a number you already track, like booked jobs, hours worked, or days to get paid?
Sources: AWS, AI readiness checklist for SMBs (September 1, 2025)
Area 2: Your information
AI that works from wrong records gives wrong answers, and it gives them confidently. AWS suggests a simple data health check: list your source systems (accounting, CRM, ecommerce, ticketing), who owns each, where the system of record lives, how data is accessed, and how often it is updated. That is an afternoon of work, and it is the most useful thing you can do before any AI project that touches customer, job, or money records.
- Do you know which systems hold your customer, job, and money records, such as accounting software, a CRM, spreadsheets, and email?
- When two of them disagree, is one officially the system of record?
- Does each customer have one record rather than several duplicates?
- Do you know who can see and change that information?
Sources: AWS, AI readiness checklist for SMBs (September 1, 2025)
Area 3: How the work gets done
AI can only follow a process that someone has defined. If the steps live in one person's head, that person has to write them down first, including the exceptions: the angry customer, the unusually large job, the insurance claim. AWS also recommends checking whether the tools you already use, such as QuickBooks, HubSpot, Shopify, Microsoft 365, or Google Workspace, can connect to AI add-ons or exchange data through an API, because tools that cannot connect make every project harder and more expensive.
- Is the task written down step by step, rather than living in one person's head?
- Do you know the common exceptions and what should happen in each?
- Is it clear who does each step today and who approves the result?
- Can your main tools connect to other software or export their data?
Sources: AWS, AI readiness checklist for SMBs (September 1, 2025)
Area 4: People and rules
People start using AI before the company has a rule for it. Unio Digital, a managed IT firm, reports that most small businesses that look find 15 to 30 distinct AI tools in use over a 90-day period, mostly consumer-tier. AWS says even small teams need clear, written rules covering data privacy, customer consent, and where a human must review AI output. Those rules can fit on one page. If you want a more formal structure later, NIST publishes a free, voluntary AI Risk Management Framework.
- Has one named person agreed to own the AI work and its results, with time set aside each week?
- Do the people who do the work know it is coming, and do they have a say in how it is built?
- Do you have a short written rule for what information staff may put into AI tools?
- Do you have a written list of situations where a person must approve AI output before it reaches a customer, such as pricing, legal topics, medical information, or large accounts?
Sources: Unio Digital, AI consulting cost for small business; AWS, AI readiness checklist for SMBs (September 1, 2025); NIST, AI Risk Management Framework
Area 5: Proof it is working
Most failed AI projects are not caught failing, because nobody wrote down what success looked like. AWS recommends tying a pilot to one or two business outcomes with a baseline, setting target thresholds before go-live, keeping metrics simple and reviewing them weekly, and pairing the numbers with quality checks such as spot reviews. The stop condition in the last question is our addition: deciding in advance when you will pause or end a project that is not moving the number keeps a bad pilot from turning into a permanent expense.
- Have you picked one or two numbers to judge the first project and written down today's baseline?
- Have you set a target before launch rather than after?
- Will someone spot-check AI output for mistakes every week in the early weeks?
- Do you have a written stop condition, a point at which you pause or end the project if the number does not improve?
Sources: AWS, AI readiness checklist for SMBs (September 1, 2025)
Score yourself
Add one point for each yes. The bands below are our own rule of thumb from working with owner-operated businesses, not a published benchmark. A high total can still hide a risky gap, so look at which area your no answers cluster in, especially Area 4. A business that scores 16 but has no rules for what goes into AI tools has a bigger problem than one that scores 12 with the rules in place.
| Score | Where you are | What to do next |
|---|---|---|
| 0 to 7 | Not yet | Pick one task, write it down, and sort out where its records live. Do not buy tools yet. |
| 8 to 14 | Nearly there | Fix the areas with the most no answers first, starting with an owner and written rules. Then begin one small pilot with a person approving the output. |
| 15 to 20 | Ready to start | Choose one pilot, set the target, and start with a person approving every output. Expand only after the number moves. |
Fixing the most common gaps
Most no answers have a small, concrete fix. The effort column is our own estimate and varies with the size of the business.
| Gap | First fix | Typical effort (our estimate) |
|---|---|---|
| The task lives in one person's head | Write the trigger, the steps, the exceptions, the owner, and the output on one page | An afternoon |
| Customer records are scattered or duplicated | List every system, name one as the system of record, and merge duplicates there first | A few days to a few weeks, depending on volume |
| No rules for AI use | Write one page: what information may go into AI tools, who approves, and what must be reviewed by a person | An afternoon, then review monthly |
| No owner | Name one person and protect weekly time for the work | A decision, not a project |
| No baseline | Track the number you want to improve for two weeks before changing anything | Two weeks of tracking |
Good first projects
AWS lists a website chatbot for common questions, email reply drafts and summaries, and invoice reminders or appointment-scheduling prompts as good starter projects, and suggests one front-line and one back-office use case, a 30 to 60 day pilot with a clear success metric, and human review in the early stages. Avoid custom, complex builds until a simple one has proven its value. For a home-services or professional-services business, the same logic points to faster first replies to new inquiries, drafting follow-ups on quotes that have gone quiet, and reminders for unpaid invoices.
Sources: AWS, AI readiness checklist for SMBs (September 1, 2025)
Check what you say about AI
Readiness also covers what you tell customers. In September 2024 the Federal Trade Commission announced enforcement actions against deceptive AI claims and said that existing consumer protection laws apply to AI. In practice, avoid calling a service "AI-powered" or promising a result unless you can show evidence for it. The same discipline applies to your vendors: ask them for evidence, not adjectives. This is general information, not legal advice.
Sources: FTC, crackdown on deceptive AI claims and schemes (September 2024)
A 30-day starter plan
If your score is under 15, this sequence closes the most common gaps in a month. AWS describes a similar 6 to 8 week "readiness sprint" that runs from a data health check through use-case selection, integration and policies, baselines, and a pilot launch.
- Week 1: pick the one problem, and write the process down on one page, including exceptions.
- Week 2: list your systems, name the system of record, and write the one-page rule for AI use.
- Week 3: name the owner, record the baseline number, and set the target and the stop condition.
- Week 4: start a small pilot with a person approving every output, and review the number weekly.
Sources: AWS, AI readiness checklist for SMBs (September 1, 2025)
When to get help
You can do everything above yourself, and this checklist is free. Some owners prefer an outside look, especially when they cannot tell which answers are really yes. Our entry point is a fixed-fee Baseline diagnostic for $2,500 that measures where the business is losing revenue and shows all the arithmetic. If you sign a build within 30 days of the report, the fee is credited toward it. For ongoing AI leadership, our fractional CAIO retainers run $10,000 to $25,000 per month a month with a 6-month minimum.
Frequently asked questions
What does "ready for AI" mean for a small business?
It means you can name one problem to solve, find the information the work needs, assign an owner, and write rules that keep AI from making costly mistakes. It does not require a data team or a large budget.
Do I need clean data before I use AI?
Not for everything. The more a task depends on your records, such as quotes, invoices, or customer history, the more those records need to be correct first. Start with tasks that do not depend on messy records and fix the records in parallel.
Do I need a policy for how employees use AI?
Yes, and it can be one page. AWS advises even small teams to write clear rules for data privacy, customer consent, and where a human must review AI output, and one managed IT firm reports that small businesses often find many AI tools already in use.
What is the difference between this checklist and a diagnostic?
The checklist is a self-check of your readiness. A diagnostic measures your actual numbers to find where money is being lost. Ours is a fixed-fee $2,500 Baseline that shows all of its arithmetic.
What should my first AI project be?
Pick repetitive, high-volume work that is clearly defined, such as faster replies to new inquiries, follow-up drafts, or invoice reminders. Run it as a 30 to 60 day pilot with one success number and a person reviewing the output.
Related reading
- What Is a Fractional CAIO? The Plain-English Guide for Business Owners A fractional Chief AI Officer is a part-time executive who decides where AI belongs in your business, sets the rules, and makes sure it does real work. This guide covers the cost, the first 90 days, the rules that apply, and who should not hire one.
- How Much Does a Fractional CAIO Cost? Every Figure, With Its Source Fractional CAIO monthly rates, full-time salary bands, staffing-firm pricing, and AI build costs for small businesses, side by side, with the source named for every number.
- Governed AI for Operating Businesses: Accountability Before Autonomy Governed AI means automation that cites verified truth, respects permission boundaries, and reports to leadership in plain language. Airus Advisory builds that operating layer for service businesses that cannot afford silent failures.
- AI Consulting Options Compared: Big Firm, Boutique, Fractional CAIO, Staffing, or Offshore Five ways to buy AI help, what each costs according to published sources, who each one fits, and what to watch for. Written for owner-operated and mid-sized businesses.