Is your business ready for AI? A simple 3-step check

Everyone is being told to "use AI." Far fewer are told what that requires. After working with operators across logistics, retail, education and healthcare, we can save you some time: AI readiness is not about technology. It's about whether your business's information is in a usable state.

AI readiness is about your data, not the technology

The tools are the easy part. The models are good, cheap and getting better every month. What decides whether AI works for you is everything underneath it — where your information lives, whether it's captured, and whether it agrees with itself. A business that's ready for AI is simply a business whose work is written down in systems instead of trapped in people's heads.

Here's the simple, three-step check.

Step one: where do your decisions live?

If prices, schedules or approvals live in someone's head or a WhatsApp thread, AI has nothing to work with yet. An agent can't act on a rule it can't see, and it can't learn from a decision that was never recorded.

The fix isn't AI — it's capturing the work where it happens, in a system. When an order is taken, a price is agreed or an approval is given, it should land somewhere structured, not just in a conversation. This step has value on its own: it makes the business less dependent on any one person's memory, AI or not.

Step two: do your numbers agree?

If two reports about the same thing give two answers, an AI built on top will be confidently wrong — and confidently wrong is worse than no answer at all, because people act on it.

Bring your data into one clean place first: duplicates merged, names matched, sources reconciled. This step pays for itself even if you never go further — every report agreeing is worth real money in saved arguments and faster decisions. It's also the foundation that makes everything after it trustworthy.

Step three: now automate

With work captured and data clean, AI agents finally have solid ground to stand on. Now automation works, because the agent is reading reliable information and following rules that actually exist in a system.

Start with one routine task with clear rules — billing, updates, reminders — and keep a person approving anything sensitive. Prove it on that single task, then expand. Because the foundation is solid, each new agent is easier and safer than the last.

The order matters

This is the part most businesses get wrong. Companies that skip straight to step three get impressive demos that quietly stop being used a month later — because the data underneath them was never ready.

Companies that follow the order get something better: a business where the routine work runs itself, on a foundation they can trust. The full journey is usually a matter of months — and the first results arrive in weeks, not years. If you're not sure which step you're on, start at step one and be honest about the answer. Most businesses are less ready than they think, and far closer to fixing it than they fear.