How to actually get your business ready for AI: a practical checklist

Most AI projects in small and mid-sized businesses don't fail because the technology is bad. They fail because the business wasn't ready to absorb it. The tools work. The pilot demos are convincing. Then the rollout meets a real organisation - messy data, unclear ownership, nervous staff, vague success metrics - and quietly stalls. Getting ready isn't glamorous, but it's the single biggest predictor of whether AI pays back for you. Here's the plain-English version of what to put in place, in roughly the order it matters.

Start with a clear, honest 'why'

Before you touch a tool, write down in two or three sentences why your business is doing this. Not 'because everyone is talking about AI'. Something like 'we want to free up ten hours a week across the ops team so they can spend more time with our top twenty clients', or 'we want to respond to inbound enquiries within an hour, every hour, without hiring'. A specific 'why' kills bad projects early and gives the team something to point at when the rollout gets hard. A vague 'why' produces a vague programme that nobody can defend at the next board meeting.

Name a single owner

AI cannot belong to 'the leadership team' or 'IT'. It needs one named person whose job it is to make this work. They don't have to be technical. They do need authority to change workflows, the trust of the leadership team, and enough time in their week to actually do the job. In most SMBs the right person is an operator - a head of operations, a strong office manager, a hands-on COO - not the IT lead and not the CEO. Without a single owner, decisions drift, pilots stretch, and accountability evaporates.

Tidy your data, just enough

You don't need a data warehouse. You don't need a six-month cleansing project. You do need to know where your core information lives, who owns it, and roughly how clean it is. Pick the two or three sources that matter most - usually customer records, sales pipeline, and a finance system - and spend a fortnight standardising the obvious problems: duplicate records, inconsistent fields, missing emails, ten different spellings of the same supplier name. Good enough is good enough. Perfect data is a fantasy and a procrastination tool.

Write a one-page acceptable-use policy

Before any tool gets rolled out, every employee needs to know what they can and cannot do with it. One page. Plain English. Cover three things: what data must never be pasted into a public AI tool (client information, financials, anything covered by NDAs), which tools are approved for work use, and what to do if something goes wrong. Don't outsource this to a law firm and don't make it twenty pages. The point is that people read it and remember it. A policy nobody has read is worse than no policy at all, because it gives leadership false comfort.

Audit what's already in use

Most SMBs we work with discover, when they look, that their team is already using three or four AI tools - some paid, some free, some on personal accounts. Map this before you buy anything new. You'll usually find duplicate spend, data going to places it shouldn't, and at least one team quietly doing something genuinely clever that the rest of the business could learn from. The audit takes a couple of afternoons and routinely pays for itself in cancelled subscriptions.

Pick the right first project

The first AI project should be small, visible, and almost boring. One workflow. One team. One owner. One success metric agreed on day one. Pick something that costs your team obvious time every week - drafting follow-up emails, summarising meeting notes, cleaning up a recurring report, triaging inbound enquiries. Avoid anything that touches sensitive customer data, anything where mistakes are expensive to fix, and anything that requires three departments to agree before it can launch. The goal of the first project is not transformation. It's to learn how your business absorbs new technology and to give your team a small, confidence-building win.

Build a small group of curious people

You don't need to train the whole company on AI. You do need three or four people who are actively curious, given permission to experiment, and connected to each other. They become the internal champions, the ones who answer 'how do I use this?' questions, and the ones who spot the next opportunity. In our experience, this group rarely needs to be recruited - they're already volunteering. Your job is to give them air cover, a small budget, and a regular slot on the leadership agenda.

Sort out the basics of governance

You don't need an AI ethics board. You do need to be able to answer four questions: who decides which tools we use, who reviews them before they go live, what happens if something goes wrong, and how do we keep a basic record of what AI is being used for. A simple register in a shared spreadsheet, reviewed quarterly, is usually enough at SMB scale. Regulators in 2026 are increasingly looking for evidence that you've thought about this - not for elaborate frameworks. Document the thinking, keep it short, keep it current.

Plan for the change, not just the tool

Buying the tool is the easy part. The hard part is the human side: people who feel threatened, people who feel sceptical, people who are quietly worried about their job. Address it early and honestly. Be clear about what AI is being used for and what it is not. Involve the team in choosing tools where possible. Show, with real examples, where AI has freed up time for more interesting work rather than replaced it. Businesses that skip this step get adoption rates of 20% and call AI a failure. Businesses that do it well get adoption rates of 80% and quietly compound an advantage.

Set a sensible budget and a review date

Pick a number you can live with for the next twelve months without needing extra board approval. For most SMBs that's somewhere between £5,000 and £50,000, depending on size and ambition. Spend it across tools, training, and a small amount of external help if needed. Review every quarter. Cut what isn't working. Double down on what is. Treat AI like any other operating expense: it needs to earn its place on the P&L.

Don't skip the basics to chase a moonshot

It is tempting, especially after a good vendor demo, to leap past the boring readiness work and aim for something headline-grabbing. Resist. The businesses pulling ahead with AI right now are not the ones running the most ambitious projects. They're the ones with tidy data, a named owner, a written policy, a curious team, and the discipline to run small pilots well. That stack of unglamorous capabilities is what turns AI from a line item into a genuine operating advantage.

A 30-day starting plan

If you do nothing else this month, do these five things. Week one: write your two-sentence 'why' and name your owner. Week two: audit the AI tools already in use and tidy your three most important data sources. Week three: draft and circulate the one-page acceptable-use policy and pick your first project. Week four: kick off the project with one team, one metric, and a thirty-day review in the diary. That's it. No consultants required, no platform purchases, no organisational restructure. Just the basics, done in the right order.

Get ready properly and the first AI project becomes a stepping stone instead of a cautionary tale. Skip the readiness work and you'll spend the next year explaining to the board why the technology that's transforming everyone else's business hasn't quite landed in yours. The technology is genuinely ready. The question is whether your business is.