The SMB Owner's AI Readiness Checklist: 10 Questions, 5 Minutes

Before you hire an AI consultancy, buy a tool, or even book a discovery call, it’s worth five minutes of your own time to work out whether AI is the right next move for your business at all.

Most SMB owners we speak to are in one of three buckets:

1. Ready now. Clear processes, digital data, a specific pain point. AI will pay back inside 90 days.
2. Close but not yet. Process chaos or data sitting in too many places. Needs a tidy-up first.
3. Not yet. The next-best pound is better spent somewhere else — website, SEO, bookkeeping, hiring.

This checklist tells you which bucket you’re in. Ten questions, five minutes, no sales call required.

1. Do you have a process that happens every day or every week that somebody on your team finds tedious?

If no: AI isn’t the problem to solve first. Find the actual pain point.

If yes: you’ve got a candidate. Write down the process in one paragraph — start to finish, who does what.

2. Is the data involved in that process mostly digital?

By “digital” I mean: emails, PDFs, spreadsheets, or data already sitting in a software system. Handwritten notes, paper-only files, and “it’s all in Steve’s head” don’t count.

If no: fix the data before you think about AI. Getting your current processes off paper and into any digital format — even a shared Google Sheet — is step one. AI built on top of a chaotic paper workflow just makes the chaos faster.

3. Can you describe, in one paragraph, what “done right” looks like for that process?

If no: you’re not ready to scope an AI project yet. Good scoping starts with a clear success definition. “We want AI to help with invoicing” isn’t scope. “We want every supplier invoice to be matched against its PO and flagged if any line doesn’t reconcile, within 24 hours of receipt” is scope.

4. Is there a single person who owns this process end to end?

If no: the project will stall on handovers. Pick someone. If the process genuinely crosses multiple roles, pick the senior person and make them accountable.

5. How many hours per week does the current process take?

You don’t need to be exact. “About a day a week” or “one person two mornings a week” is enough.

This is how you judge return on investment later. A £4,000 AI project that saves 5 hours a week pays back in the same quarter at any reasonable labour cost.

6. What systems does the process touch?

Rough list. “We use Sage for accounting, Outlook for email, and a shared drive for PDFs.” The more mainstream the systems (Sage, Xero, QuickBooks, Google Workspace, Microsoft 365, Shopify, HubSpot), the easier and cheaper an AI project tends to be. Very niche or legacy systems add cost.

If the answer is “WhatsApp and memory”: that’s the problem, not AI. Fix the system layer first.

7. Do you already have 10+ real examples of the input data you’d be feeding the AI?

This matters more than people think. An AI system for reading invoices needs real invoices to be trained, tuned, and tested. A customer enquiry triage agent needs real customer enquiries.

If you can put 10 or more examples in a folder right now, you’re ready. If not, the first step might just be collecting them.

8. What happens right now when the process goes wrong?

If the answer is “we notice within the week”: fine, most automation pilots should include a human-in-the-loop review stage.

If the answer is “we don’t notice for a month”: you’ve got a bigger problem than AI can solve. Fix your oversight first, because an AI system that runs without proper checks can create compounding errors fast.

9. Are you prepared to involve your team in the build?

Applied AI projects that succeed include the people currently doing the job. They know the edge cases, they know the tricks, they know what a “good” output actually looks like. Projects that fail treat the team as something to work around.

If the answer is “the team will resist any change”: your first conversation isn’t with an AI consultancy. It’s with your team, about why they’d welcome this.

10. Is your website and digital foundation strong enough that AI won’t be holding up a house built on sand?

This is the one most SMB owners skip. If your customers can’t find you on Google, if your website doesn’t convert visitors, if your enquiry form has been broken for six months — AI isn’t the next priority. A working website is.

We’ve told plenty of prospective clients exactly this: spend the money on the website first, come back for AI when the front door is working.

How to score yourself

• 8 or more yes answers: You’re ready. A scoping audit is worth your time.
• 5 to 7 yes answers: You’re close. Identify the gaps and fix one or two before committing to a build.
• 4 or fewer yes answers: Not yet. The bigger win is somewhere else — usually process tidying, website, or SEO.

What to do next

If the checklist says you’re ready:

1. Pick the single most painful process from question 1.
2. Write the one-paragraph description of “done right” from question 3.
3. Pull 10+ real input examples into a folder.
4. Book a 15-minute intro call.

If the checklist says you’re close:

1. Fix the specific gap (usually data or ownership).
2. Re-score in 30 days.

If the checklist says not yet:

1. Don’t spend money on AI. Spend it on whatever came up most often — website, process, hiring, or SEO.
2. Come back to this checklist in six months.

Frequently asked questions

Is this checklist biased toward selling us something? Genuinely no. Every question exists because we’ve seen projects fail when the answer was “no” and the client was talked into proceeding anyway. The checklist exists to stop you wasting money.

What if we’re ready for some processes but not others? Pretty common. Pick the most ready one. You don’t need to automate the whole business to get meaningful return.

How do we know if our systems are “AI-friendly”? As a rule of thumb: if your system has an API, or a clean export to CSV, or email notifications on events, it’s AI-friendly. Mainstream UK business tools (Sage, Xero, QuickBooks, HubSpot, Shopify, Microsoft 365, Google Workspace) all qualify.

How long does an AI scoping audit take? Usually one to two weeks from first call to written report. You come out with a prioritised list of opportunities, rough costs, and expected paybacks — and it’s yours to keep whether you proceed with us or not.

If you scored 8+, book a scoping call. If you scored lower, use the list to fix the gaps first. Either way, you’re making a better-informed decision than 90% of businesses that say yes to AI because it sounded good in a sales deck.