Is your business ready for AI? How to tell without a consultancy

You are unsure whether AI is anything for your business. Not because you think it is nonsense, but because you have no idea whether you are ready for it. You have no data team, no IT department and no time. And the people who could tell you whether you are ready make their money by saying yes.

The good news: you can judge this yourself. You do not need a consultancy for it and you do not need an expensive baseline study. You need four questions you can answer in an hour, provided you are honest with yourself.

Being ready is not about technology

Almost everyone thinks about technology when this question comes up. Are we in the cloud, do we have modern software, is there anyone here who understands computers. That is not what decides it. We come across businesses with dated packages where it works out fine, and businesses with everything brand new where it falls apart.

What does decide it are four things that have nothing to do with servers: how you work, what you record, who wants it, and what you expect from it.

Is my business ready for AI? Four things decide the answer

Work through these four areas. Each one comes with a few questions you can genuinely answer, without having to look anything up. Answer with what is true, not with what you would like to be true.

1. Your processes: can you write down how the work goes

A system can only take over work somebody can explain. Not in broad strokes, but step by step: what comes in, who does what, when, and what the exception is. If the answer to “how does that work now?” is that it depends and that Ellen does that, the process itself is not finished yet.

Take an installation company with a shared info mailbox. The question is not whether AI can sort email. It is whether you can write down when something is a breakdown and when it is a quote request, and what happens when it is both. If nobody can, you do not get an automation project but a discussion about the way you work, with a software bill attached.

• Can you put the process you have in mind on one sheet of paper in ten steps, without anyone having to ring round to check?
• Do two colleagues who do the same work also do it the same way?
• Do you know how often it happens: per day, per week, per month?
• Can you name three exceptions that come up regularly, and what happens then?
• Is there someone who knows the whole process from start to finish, or does everyone only know their own bit?

Two or more noes does not mean you are not ready for AI. It means the first step is writing it down, not automating. That is an afternoon of work per process and it gains you something regardless.

2. Your data: is it somewhere usable, or only in heads and mailboxes

AI is not clairvoyant. Whatever a system has to use has to be somewhere it can reach: in your package, in a folder, in a list. Not in your operations manager’s head and not scattered across five personal mailboxes. In practice a surprising amount of the knowledge in an SME is recorded nowhere at all.

At a wholesaler this tends to be about product numbers. If the same product sits in the system under three spellings from three suppliers, sometimes with a space and sometimes with a hyphen, a person reads through that. A system cannot. At an accounting firm it is the same story with client names: the limited company, the holding company and the private individual are all entered slightly differently.

• Is the information that is needed held in a system, or in mailboxes, WhatsApp groups and loose Excel files?
• Do customers and products appear once in your package, or do you regularly come across duplicates?
• Can you export a list from your package, for example as an Excel or CSV file?
• Are there fields that were once meant for something else and are now full of notes?
• Do you know whether your main package has an integration, and if so, whether it is included in your subscription?

No business has perfect data. You do not have to have everything in order, only the part the first project touches. But be honest: cleaning that up is almost always the biggest part of the first weeks, and it never appears in any demo.

3. Your people: is there someone who wants to carry it

This is the quiet failure. Technology is rarely the problem; ownership is. There has to be someone in your business who suffers when the process does not work, who gets time to keep an eye on it, and who is not doing it on top of everything else in the evenings.

It does not have to be the director and it does not have to be someone who understands software. In practice it is often the work planner, the admin assistant or the scheduler: someone who knows the work and is better off for it. What does not work is a project that belongs to the director while the director is too busy. Then nothing happens, and after three months nobody mentions it again.

• Can you name right now the person who is going to carry this?
• Can that person really spend a few hours a week on it, and does that come off other work or is it added on top?
• Does that person benefit from it themselves, or are they only doing it because they have to?
• Are the people who will use it behind it, or do they hear about it at handover?
• Is there someone who understands where the controls are after handover, or does everything hang on the builder?

4. Your expectations: do you want a miracle or half an hour a day

This decides whether a project feels successful more often than the technology does. If you expect the admin workload to halve, half an hour a day is a disappointment. If you expect half an hour a day, you are satisfied with exactly the same outcome.

And do not count on saved wage costs. Remove five hours of retyping a week at a haulage firm and you do not let anyone go. That work shifts to what was left undone: chasing calls, scheduling, checking runs. Valuable, but it does not show up as a minus on your profit and loss account. The gain you can actually prove sits in work you are currently losing: quotes that go out too late, invoices left sitting for weeks, customers who walk away because they got no answer.

• Can you say in one sentence what should be different in three months?
• Is that something you can count, or is it a feeling?
• Do you already know the current number, or do you have to measure it for a few weeks first?
• What counts as a disappointing outcome for you, and would you recognise it when it happens?
• Can you live with a system that works well in nine out of ten cases and where someone still takes a quick look?

Being ready for AI does not mean everything is in order. It means you know what is not in order.

When you are better off waiting a while

Sometimes all four areas are perfectly fine and you should still not start. Not because AI is nothing for you, but because your business has no room for it right now. Postponing is a cheap decision; running aground halfway is not.

• You are in the middle of a system migration. Wait until you are running on the new package. Everything you connect to the old system now, you will build again later.
• There is an acquisition or a handover of the business under way. As long as it is not settled who decides and which systems stay, every investment in integrations is a gamble.
• You are in your busiest season. At a construction firm in spring or a wholesaler before the holidays, nobody gets hours free to keep an eye on it.
• The person who would carry it has just started or is about to leave. Wait until you know who it is.
• You genuinely need the money for something else right now. Automation rarely pays for itself within a quarter.

Waiting is not a licence to do nothing, by the way. Most of the preparation costs no money, only attention.

What you can usefully do in the meantime

The work below is always worth doing. Even if you never automate anything, your business runs better for it, and it makes you less dependent on that one colleague who knows everything.

1. Write down one process the way it really goes, not the way it is supposed to go. Follow it through once and count the actions and the waiting time.
2. Note the exceptions you come across along the way. There are always more of them than everyone thought and they decide half the work later on.
3. Clean up your most important list. Merge duplicate customers, align product numbers, complete addresses. Dull work you needed to do anyway.
4. Measure one number you want to see again later. Days between finished work and invoice sent, or hours a week spent retyping.
5. Switch on what you already have. Many packages now have an AI feature nobody has ever enabled. That costs you an afternoon and nothing more.
6. Put information that currently only lives in people’s heads on paper. Who calls which supplier, which customer has non-standard terms, why we do it this way.

If you would rather not judge those four areas on your own, our free AI scan is the tool for it: you walk through the same points and get back where your weak spot is. If the outcome is that you are better off tidying up first, it says so.

How to read your own outcome

If you score reasonably on all four, you are ready enough for one small project. Not your biggest problem, but something visible. If you score well on three and badly on one, start with that one area. If it is your data, go and clean up. If it is your owner, appoint someone or postpone.

If you score badly on three or four areas, automating is the wrong answer right now. That feels like a rejection and it is not. It is the cheapest advice you can get, because the bill for the alternative only arrives after six months.

Where this self-test falls short

To be fair: questions you answer yourself have a blind spot. You do not know what you do not know. We regularly see a director who is convinced a process “basically always goes the same way”, while on the shop floor you see six variants in two days that nobody had mentioned.

The second limitation: these questions say something about whether you are ready, not about what you should do. Green on all four does not mean there is a project worth doing. Your business can be perfectly prepared and there is simply too little repetition in your work to gain anything. That is an outcome too.

When you need us and when you do not

If you work it out yourself with these four areas, then by all means do. If you have a clear process, someone to carry it and a package that has the feature on board already, hiring anyone in is a waste of your money.

We are useful if you doubt whether your judgement is right, if the process runs across several systems that do not talk to each other on their own, or if custom software is needed that no standard package delivers. With us that starts with a core session: a half-day (four hours) on site with you, in which we follow one process from start to finish. That costs €596 excluding VAT and travel expenses. Build work after that is charged per hour: €110 for straightforward work, €165 for complex development.

And if after that half-day we think you are better off waiting a while, we say so. That is a short conversation, but it saves you a project you would not have finished anyway.