AI readiness assessment: a practical scorecard for service businesses

Most 'are you ready for AI' content is fluff. Here is an actual scorecard you can run in fifteen minutes, across the six things that decide whether an AI project pays back or dies — with what each level looks like, a score out of 12, and the right first move for your result.

Most "is your business ready for AI?" content is useless — a list of buzzwords ending in a sales pitch. So here is the opposite: an actual scorecard you can run on your own business in about fifteen minutes, across the six things that genuinely decide whether an AI project pays back or quietly dies.

The important idea up front: AI projects almost never fail on technology. The models are good enough. They fail on ambiguity — no one could say which process to fix — on data trapped in people's heads, on having no owner, or on fantasy expectations. Readiness is about those human and process factors, not about whether you have the latest tool.

The six dimensions

Score each dimension 0, 1 or 2 using the descriptions below. Be honest — an inflated score just wastes your own money later.

1. Use-case clarity

Do you actually know the one process that is costing you the most?

• 0 — "We should do something with AI" but no specific process in mind.
• 1 — A few candidates, no clear winner.
• 2 — You can name the single most expensive repetitive process and roughly what it costs in hours.

2. Process consistency

Is that work repeatable and rule-based enough for a machine to help?

• 0 — Every case is bespoke and lives in one expert's judgement.
• 1 — Mostly consistent with frequent exceptions.
• 2 — A clear, repeatable pattern most of the time (e.g. enquiry → quote, invoice → posting).

3. Data accessibility

Can the information the AI needs actually be reached?

• 0 — It is in people's heads, on paper, or scattered across personal chats.
• 1 — Digital but messy and spread across tools.
• 2 — The relevant data lives somewhere reachable — WhatsApp Business, a CRM, email, spreadsheets.

4. Tool connectivity

Do your systems allow anything to connect to them?

• 0 — Closed, offline, or paper-based tools.
• 1 — Modern tools but you are unsure about integrations.
• 2 — You use tools with APIs or integrations (Jobber, Zoho, QuickBooks, Xero, WhatsApp Business API, Google Workspace).

5. Ownership

Is there a human who will own the change and approve AI output at first?

• 0 — Nobody has time; it would be "everyone's" job.
• 1 — Someone interested but not resourced.
• 2 — A named person will own it and review the AI's work during rollout.

6. Expectations & budget

Are your expectations realistic, and will you fund a small pilot?

• 0 — Expecting AI to replace judgement entirely, or unwilling to spend anything.
• 1 — Interested but cautious and vague on budget.
• 2 — You see AI as augmenting your team and will fund a scoped pilot to test the return.

Score yourself

Add up your six scores for a total out of 12.

The three gaps we see most in Dubai SMEs

Across the businesses we assess, the same three readiness gaps come up again and again:

• Data scattered across WhatsApp and paper. The enquiries, quotes and job notes exist — but across personal phones, notebooks and memory. The fix is rarely a big "data project"; it is usually routing one channel (say, all enquiries) into one place the AI can read.
• Boil-the-ocean ambition. Owners want to automate everything at once, so nothing ships. The fix is discipline: one process, proven, before the second.
• No owner. A tool gets bought, nobody drives adoption, and it dies within a month. The fix is naming a person before you build, not after.

Notice that none of those are technology problems. They are the real content of "readiness".

A worked example

Before we built quoting automation on our own company, Fix It Mates, we ran it through exactly this scorecard. It scored well on use-case clarity (quoting was clearly the most expensive repetitive process), process consistency (enquiry → itemised quote is a repeatable pattern), data accessibility (enquiries came through WhatsApp) and tool connectivity (we quote in Jobber, which has an API). The one weak spot was ownership — at first no single person owned the rollout, which slowed it down until we fixed it. Total: high enough to pilot, and the pilot is what became the system that now drafts a quote in about a minute.

The lesson: you do not need a perfect score. You need one high-value, consistent process, data you can reach, and someone to own it. If your data is messy, that is normal — a well-scoped first project usually tidies it as a by-product. The fastest way to get a precise read on your own business is a done-for-you version of this: our free AI Operations Audit scores exactly these dimensions and hands you a prioritised plan with rough cost and payback.

Frequently asked