How to use this guide
AI tools are now cheap to try and easy to buy. That is exactly why so many businesses end up with subscriptions nobody uses and pilots that never reach daily work.
This scorecard covers the 25 things I check first before recommending any AI tool to an owner-led business. Tick each statement that is true today. Every tick is a gap that will make AI harder, slower or riskier.
A tip from experience: be strict with section 1. If you cannot name the problem and the number you want to move, no tool will fix that for you.
1A clear business problem
1. You want to "use AI" but cannot name the specific task it should improve.
Tools chosen before the problem is defined rarely make it into daily work.
First fix: Write down the one task that costs the most time or causes the most errors, and start there.
2. You have not measured how long that task takes today.
Without a baseline, you cannot tell whether AI helped or simply added work.
First fix: Time the task for one week: how often it happens, how long it takes, how often it goes wrong.
3. Nobody has agreed what success would look like.
Pilots without a target drift on until everyone quietly stops using them.
First fix: Agree one measurable result, for example "reply to enquiries within one hour" or "halve time spent on quotes".
4. The task changes every time, with no repeatable pattern.
AI works best on tasks that follow a recognisable pattern with clear inputs and outputs.
First fix: Pick a task with a clear structure first. Leave the judgement-heavy work for later.
5. You are buying AI because competitors or suppliers say you should.
Pressure-led purchases solve someone else's problem, not yours.
First fix: Before any purchase, write one sentence on the business result it must deliver, and check it after 60 days.
Related service: CRM, Automation and AI
2Processes that are written down
6. Key processes live in people's heads rather than on paper.
You cannot automate or improve a process nobody can describe consistently.
First fix: Write the steps of your top three processes on one page each, in plain language.
7. Different people do the same task in different ways.
AI amplifies whatever process it is given, including inconsistency.
First fix: Agree the best way to do the task and make it the standard before adding any tool.
8. Hand-offs between people or teams regularly cause delays.
Delays at hand-offs are often a process problem, and better solved by routing than by AI.
First fix: Map who hands what to whom, and fix the slowest hand-off first.
9. There are no templates for common documents, messages or quotes.
Templates are the simplest form of automation and the foundation for AI drafting.
First fix: Create templates for your five most common documents or replies.
10. Nobody reviews how work actually flows from enquiry to invoice.
Without the end-to-end view, AI gets bolted onto the wrong step.
First fix: Walk one recent job from first contact to payment and note every step and delay.
Related service: Business Operations and Commercial Performance
3Data you can trust and reach
11. Customer and job information is spread across email, WhatsApp, spreadsheets and paper.
AI can only work with information it can reach. Scattered data means scattered results.
First fix: Bring customer and job records into one system, such as a CRM, before adding AI on top.
12. The same customer appears several times with different details.
Duplicate and outdated records lead to wrong answers delivered confidently.
First fix: Clean your main customer list: merge duplicates and remove records you no longer need.
13. You are not sure which data is sensitive or confidential.
Sensitive data put into the wrong tool is a privacy and trust risk.
First fix: List the types of information you hold and mark which are personal, confidential or regulated.
14. Reports take hours to prepare by hand.
If people cannot see the numbers easily, AI insights will not be trusted either.
First fix: Set up one simple weekly report from your main system before asking AI to analyse anything.
15. You do not know who has access to which systems and data.
Unclear access becomes a bigger risk when AI tools can read and act on that data.
First fix: List every system, who has access, and remove access that is no longer needed.
Related service: CRM, Automation and AI
Prefer to tick it on paper?
Get the printable version with tick boxes, section subtotals, the scoring guide and a worksheet. You also get the rest of the EH Advisory library and one practical fix a month; unsubscribe any time.
4People and skills
16. Staff use AI tools on their own, with no shared guidance.
Unguided use leads to inconsistent quality and information shared where it should not be.
First fix: Agree a one-page guide: which tools are approved, what may and may not be shared, and who to ask.
17. Nobody owns AI and automation in the business.
Without an owner, pilots stall and nobody learns from what worked.
First fix: Name one person to own AI pilots, with an hour a week set aside for it.
18. The team has not been asked where they lose the most time.
The people doing the work usually know exactly where the waste is.
First fix: Ask each person for the one task they would most like to stop doing by hand.
19. There is no time set aside to learn and test new tools.
Tools adopted in a rush are dropped in a rush.
First fix: Run a small, time-boxed test of 30 days on one task, with one person responsible.
20. People worry AI will replace them, and nobody has addressed it.
Worried teams resist change quietly, which stalls even good projects.
First fix: Be explicit about the goal: removing repetitive work so people can spend time on customers and judgement.
Related service: CRM, Automation and AI
5Risk, privacy and governance
21. You have not checked where your AI tools store and process data.
Many tools process data in other countries, which brings privacy obligations.
First fix: For each tool, check where data is stored, whether it is used for training, and how to switch that off.
22. AI-generated content goes to customers without a human check.
Confident but wrong answers damage trust faster than slow ones.
First fix: Require a human review for anything that goes to a customer, until accuracy is proven on your own work.
23. There is no record of which decisions AI helped to make.
If a decision is challenged, you need to show how it was made.
First fix: Keep a simple log of where AI is used in decisions that affect customers, staff or money.
24. Your privacy notice does not mention how you use AI or automated tools.
Customers and regulators increasingly expect to be told.
First fix: Update your privacy notice to describe the tools you use and how personal information is protected.
25. You have not considered what happens if a key AI tool stops working or changes price.
Dependence on one tool without a fallback is an operational risk.
First fix: For each critical tool, note the manual fallback and an alternative supplier.
Related service: Governance and Regulated Operations
What your score means
Ready to pilot
Your foundations are in place. Choose one well-defined task, set a target and run a 30-day pilot.
Fix the foundations first
Where most owner-led businesses land. Tighten processes and data for 4 to 8 weeks, then pilot.
Not ready yet, and that is useful to know
Buying tools now is likely to waste money. Start with sections 1 and 2.
Run one small, safe pilot
Pick the task with the highest frequency and the lowest risk. Set one target, give one person 30 days, review the result, and only then decide whether to scale or stop.
Want a second pair of eyes on your score?
Book a 30-minute call. Bring your ticks and we will look at your biggest gap and what it would take to close it.
Questions
What is an AI readiness assessment?
A structured look at whether a business has a clear problem, documented processes, usable data, the right people and sensible controls in place before it invests in AI tools.
How do I know if my small business is ready for AI?
If you can name the task you want to improve, measure it today, and the data it needs is in one place and reasonably clean, you are ready to run a small pilot.
Who is this scorecard for?
Owner-led service, professional, financial services and hospitality businesses considering AI tools, chatbots or automation.
What is in the printable version?
Tick boxes, section subtotals, the scoring guide and a worksheet to choose your first AI use case.
This guide is general guidance, not financial, tax or legal advice.