Your business is ready for AI when a specific expensive problem meets usable data and a human who owns the process. That's the whole test. The technology is rarely the blocker — a vague problem, missing data, or nobody accountable is. Score yourself honestly against the ten signals below and you'll know exactly where you stand.
We built this the way we'd assess it ourselves before quoting a project. It is deliberately hard to pass by fluke. A high score means go; a low score means you'll waste money if you start now — and that's worth knowing before you spend it.
How do I know if my business is ready for AI?
Work through ten yes/no signals and count the yeses. Each one is a real precondition for an AI project paying off. Answer honestly — a generous self-score just moves the disappointment from now to later, after you've spent money.
Here are the ten. Score one point per genuine yes.
| # | Signal | You can say yes if… |
|---|---|---|
| 1 | A specific, expensive problem | You can name one task and put a number on what it costs in hours or lost revenue |
| 2 | The data or content exists | The information the task needs is written down somewhere a tool could read |
| 3 | The data is usable | It's reasonably consistent and current, not scattered across heads and sticky notes |
| 4 | A human owns the process | One named person is accountable for the task today and would own the AI version |
| 5 | Appetite for change | The team is willing to change how they work, rather than bolt a tool onto old habits |
| 6 | You can measure a result | There's a before-number you could compare an after-number against |
| 7 | Realistic budget | You can fund a small first step without betting the business on it |
| 8 | Repetitive volume | The task happens often enough that saving time on it actually adds up |
| 9 | Tolerance for a wrong answer | A mistake can be caught by a human before it does real damage |
| 10 | Leadership backing | Someone with authority wants this to work and will unblock it |
Don't rush. The value of this list is entirely in how honestly you answer it.
What does each signal really mean?
Each signal maps to a way AI projects fail in the real world. Understanding the failure it guards against makes it easier to score yourself straight. Below is what sits behind the ten.
- A specific, expensive problem. Projects that start from "we should do something with AI" almost always fail, because there's no target to hit. A costed problem gives you one.
- The data or content exists. AI reads and reasons over information. If the knowledge lives only in a colleague's head, there's nothing for a tool to work with yet.
- The data is usable. Existing-but-contradictory data is worse than none — the model confidently repeats your mess. This is the most common quiet killer of AI projects.
- A human owns the process. Tools drift and degrade without an owner. Someone has to check outputs, spot when it's wrong, and keep it honest.
- Appetite for change. AI usually changes the workflow itself, not merely the tooling around it. Teams that want to keep every old step and add a robot on top get the cost and none of the gain.
- You can measure a result. No baseline means no way to tell if it worked, which means no way to justify the next step. Measurement is what turns a pilot into a decision.
- Realistic budget. Ready doesn't mean rich. It means you can fund a small first step — free tools then, maybe, a few thousand pounds — without staking the firm.
- Repetitive volume. A five-minute saving on a once-a-year task is noise. The same saving on a daily task is a part-time hire's worth of time back.
- Tolerance for a wrong answer. Good first projects are forgiving — a human catches errors before they reach a customer. If a single wrong output is catastrophic, that's a later, more careful project.
- Leadership backing. Someone has to clear diaries, approve access, and defend the experiment when it's mid-flight and not yet proven. Without cover, pilots stall.
What does my score mean?
Add up your yeses. The band you land in tells you whether to build, test, or fix the groundwork — and roughly what to spend. This is a genuine read, not a sales funnel dressed as a quiz.
| Score | Where you stand | What to do |
|---|---|---|
| 8–10 | Ready to move | Run a cheap experiment now, then scope a real project. You have the conditions to succeed. |
| 5–7 | Ready to test, not to build | Start on free or low-cost tools against your best problem. Fix the missing signals before spending on a build. |
| 3–4 | Groundwork first | You've likely got a problem but not the data, owner, or measurement. Sort those before any AI spend. |
| 0–2 | Not yet — and that's fine | AI isn't your next move. A process fix, better data capture, or clearer ownership will pay back faster. |
Notice what the bands are really measuring. A low score almost never means "AI can't help you." It means the preconditions aren't in place, and spending on AI before they are is how money gets wasted. Nearly every low score is a short, fixable list.
What do I do at each score?
Match your action to your band, and resist the urge to skip ahead. The most expensive mistake is a 4-scorer commissioning a build that a 4-scorer isn't ready to use well.
If you scored 8–10 — move, but stay disciplined. You have the conditions. Now don't waste them by boiling the ocean. Pick the single best problem, run a two-week cheap test to confirm the signal, then scope one real project. Our how to start using AI guide walks the exact 90-day path. When you're ready to size a real build, a fixed-fee AI System Audit de-risks it before you commit build money.
If you scored 5–7 — test before you build. You're close. The missing signals are usually data quality or measurement. Prove value first on free or low-cost tools, keep the results, and use them to justify — or reject — a bigger spend. Testing cheaply is exactly how you earn your way into the 8–10 band with evidence instead of hope.
If you scored 3–4 — fix the groundwork. You can likely name a problem, but the machinery around it isn't there. Common fixes: appoint an owner for the process, start writing down knowledge that only lives in people's heads, and set up a simple before-measurement. None of this needs AI. It's the unglamorous work that makes AI pay off later. The 7 mistakes UK businesses make guide is mostly a list of what happens when this step gets skipped.
If you scored 0–2 — do something else first. This is a real and honest result. AI is not your next best pound. A tighter process, a form that captures data properly, a clear owner for the messy task — one of those will beat AI for you right now. Sometimes the answer to "is AI the wrong tool?" is simply yes, for now — and we cover exactly when in when AI is the wrong tool.
Why we tell people they're not ready
Because a badly-timed AI project is worse than none. It burns budget, sours the team on the whole idea, and leaves you convinced "AI doesn't work for us" when the truth is the timing or the groundwork was off. We'd rather send you away with a free tool and a to-do list than take money for a build you can't yet use.
That's not modesty — it's how the numbers work. A business that fixes its data and appoints an owner then builds gets a system that works. A business that skips those and builds anyway gets an expensive thing gathering dust. The second one doesn't come back, and doesn't refer us. Honest readiness advice is just good business on both sides.
There's also a regulatory reason to be straight about readiness. If your problem touches personal data or high-stakes decisions, UK GDPR and the EU AI Act shape what you can responsibly build. "Ready" includes knowing those constraints before you spend, not discovering them after.
How does this map to the free Readiness Assessment?
The ten signals above are the shape of the conversation we have in the free AI Readiness Assessment. The difference is we do it with you, against your actual business, and we'll dig into the ones you scored generously. It's free by design — the whole point is an honest read before anyone spends.
If you scored 5 or above and want that second opinion, it's the natural next step. If you scored below 5, you probably already know what the groundwork is — the assessment just helps you sequence it so AI lands well when you're ready.
FAQ
Is my business ready for AI? You're ready when three things line up: a specific expensive problem, data or content the task can actually read, and a human who owns the process and can spare time to run it. If any one is missing, fix that first. Readiness is about the problem and the people, not the technology.
What is an AI readiness checklist? A short self-assessment of the conditions that make an AI project succeed — a clear costed problem, usable data, process ownership, appetite for change, budget realism, and a way to measure results. Scoring yourself honestly tells you whether to start now, start small, or fix the groundwork first.
Do I need clean data before using AI? For a first experiment, no — the data needs to exist somewhere readable, not be perfect. For a production build that makes decisions, data quality matters a great deal. Messy, contradictory, or missing data is the single most common reason AI projects underdeliver.
How do I know if AI is worth it for my business? Put a number on the problem first. If a task costs you real hours or lost revenue every week, and you can measure it before and after, AI is worth testing. If you can't cost the problem or measure a result, it isn't worth spending on yet.
What if my business isn't ready for AI? That's a useful answer, not a failure. Usually one or two fixable things are missing — an owner, a written-down process, or a clearer problem. Sort those, run a cheap experiment, and reassess. Not-ready-yet is almost always a short list of groundwork, not a permanent no.
Does my business need a data scientist to be AI-ready? No. Readiness is about a costed problem, usable data, and process ownership. Applying AI can be done with off-the-shelf tools or a small build. Specialist hiring is a later-stage question once value and volume are proven.
How much should a business budget to get started with AI? Start near zero — free or low-cost tools to prove the problem. Only climb to a paid tool, an automation (a few thousand pounds), or a build (from around £12,500) once a cheap test shows a real saving. Readiness includes being realistic that the first step should be small.