What an AI readiness assessment is (and is not)
An AI readiness assessment is a structured look at whether your business can use AI tools safely and usefully, judged on what is actually happening inside it today. Done well, it tells you which tasks fill your people's week, which tools they already use (approved or not), what rules exist and what is missing before you spend money on anything new.
The term covers a lot of ground. It can mean a free online quiz, a vendor's discovery workshop or a six-week consulting engagement. They answer different questions. A quiz tells you how confident you feel. A workshop tells you what the vendor sells. What an owner of a 15 to 150 person business usually needs is plainer: where do we stand right now, and what should we do first?
It is not a certification, a compliance sign-off or a business case for a particular product. If an offering claims to be any of those, read the fine print.
Start with what people already do
Most readiness conversations begin with 'what should we buy?' The more useful starting point is 'what are we already doing?' In many businesses the answer is more than the owner expects. Someone in the front office pastes client emails into a free chatbot to tidy the wording. A project coordinator has an AI notetaker joining Teams calls. A salesperson installed a browser extension that summarises web pages, and it can read every page they open.
None of this is sinister. These are people trying to get through a heavy week. But it means the business is already using AI, just without a decision behind it. An assessment that only asks leadership about plans will measure the business you intend to have and miss the one you run.
So the first job of an assessment is to find out what is true: which tools are in use, on which accounts, for which tasks, and with what kind of information going in.
Picture a ten-person accounting firm. The partner approves one tool and believes that is the whole story. A short survey then finds reception using a different chatbot for client letters, a graduate running an AI notetaker at client meetings, and the bookkeeper using a spreadsheet add-in nobody has heard of. The policy question changes completely once you know that.
The six areas a good assessment covers
Providers group things differently, but a useful assessment touches six areas. You can use this list to test any assessment you are offered.
- Strategy and purpose: what the business is trying to get better at, and who decides. 'Use AI' is not a purpose. 'Stop re-keying supplier quotes by hand' is.
- Process and use cases: the repetitive tasks that fill the week, such as minute-taking, first-draft correspondence, data entry between systems or chasing documents, and which of them are realistic candidates.
- People and skills: who is confident, who is anxious, who has been trained, and who has quietly become the office expert.
- Data and technology: where information lives, how tidy it is, which systems people work in, and whether sensitive material sits somewhere an AI tool could reach it.
- Governance and risk: whether there is a written policy, who approved the tools in use, and what happens when something goes wrong.
- Insurance and cyber: the controls clients and insurers tend to ask about, such as multi-factor authentication, access reviews and whether anyone has read your policy wording with AI in mind.
Who should be asked, and how
Ask everyone, not just the leadership team. Owners and managers know what they have approved. Staff know what they actually do. The gap between those two is usually the most valuable finding in the whole exercise.
Keep it short. Eight minutes on a phone gets answers from people who would never book a workshop slot. Ask each person about their own role, the tasks that fill their week, the tools they use and how they feel about change, and let them answer without a manager reading over their shoulder.
Anonymity matters more than owners expect. If staff think the answers will be used to decide who is needed, they will describe a tidier, more cautious workplace than the real one. Report results in groups so no individual can be picked out (three people is a sensible minimum) and say plainly what the results will and will not be used for.
Ask owners and managers a few separate questions about policy, systems and insurance. Those answers sit differently from staff experience and should be reported separately.
Think about who is not at the table. Part-time staff, people on the road, site-based crews and anyone without a desk are the most likely to use tools on personal phones. If the method only reaches people at a computer, the picture will skew toward the office.
What you should have in your hands at the end
Be wary of scores in particular. A number out of 100 is only useful if you can see what drove it and compare it with your own previous result. A score that means something different every time is decoration.
The test of a result is whether it changes what you do. If you finish reading and your next step is still 'form a working group', the assessment has not done its job. You should be able to point to one or two things that happen next week.
A good assessment ends with decisions you can make on Monday, not a document to file. Look for these:
- A summary of where you stand in each area, with scores or ratings only if the basis for them is explained.
- A tool register: the tools people say they use, including personal accounts and free versions, and for which tasks.
- A controls checklist covering the questions clients and insurers tend to raise.
- A prioritised list of recommendations, each with a plain next step.
- Planning estimates of where hours go on repetitive work, framed as figures to test rather than promised savings.
- A 30, 60 and 90 day plan, so the first quarter has an order to it.
Signs it is a sales pitch in disguise
A few patterns are worth watching for. Ask who wrote the questions, who sees the raw answers and what happens to the data afterwards. A good provider answers each in a sentence.
- The assessment is free, but the output is a proposal for the provider's own services.
- It is based on a workshop with managers and has no way of seeing what staff actually do.
- Scores arrive without any explanation of how they were reached.
- It promises percentage savings or productivity gains before anyone has looked at your tasks.
- It uses certification language such as 'AI compliant' or 'audit passed'.
Ask: if someone in our front office used a personal AI account for client work last week, would this assessment have found it? If the answer is no, it is measuring intentions, not reality.
When it is worth doing
An assessment earns its keep when at least one of these is true: staff are probably already using AI and you have no policy; a client or insurer has started asking questions; you are about to buy a tool and want to know where it would genuinely help; or you tried something and it stalled.
It is not worth doing as a box-ticking exercise, and it is less useful if nobody on the leadership team will act on the result. Pick one person to own the output before you start.
Calibre Forum is one way to get this: a short survey each person answers on their phone, with anonymous group results and a report for the owner afterwards. Doing it yourself with a spreadsheet and a round of honest conversations can also work, if you are disciplined about asking everyone and reporting what you hear.
Whichever route you take, treat the first assessment as a baseline. Run a lighter version again in a few months and see what moved.
Timing matters. A good moment is before a renewal, before a significant software purchase, or before you write a policy, because the result changes what you do next. A poor moment is the week of a restructure, when every question will feel loaded and people will answer defensively.
Common questions
How much does an AI readiness assessment cost?
It depends on the format. Free online quizzes cost nothing but tell you little about your own business. Consultant-led engagements are usually priced by effort and can run to many thousands of dollars. Whatever you consider, ask for a fixed price and a written list of exactly what you will receive at the end.
How long does an AI readiness assessment take?
A few weeks at most, and most of that is waiting for people to respond. A staff survey should take each person under ten minutes. The owner's effort is a short setup and then reading and acting on the results. Anything that needs weeks of workshops is probably built around the provider's calendar, not yours.
Do we need to be using AI already to do an assessment?
No. It is often more useful if you think you are not, because you may be wrong. People commonly use AI tools without telling anyone. An assessment finds out whether that is happening, and if it is not, it shows what a sensible first step looks like for your tasks and your size.
Is an AI readiness assessment the same as an AI audit?
No. An assessment gives you a picture of current use, gaps and priorities so you can decide what to do. An audit usually tests specific controls against a defined standard and produces a formal opinion. An assessment is also not a certification, and it is not legal, tax or insurance advice.
This guide is general information, not legal, insurance or financial advice. Check your own obligations with a qualified adviser.