Why we tell some businesses not to use AI yet

This might seem like an odd thing for an AI consultancy to say, but we regularly tell businesses not to use AI. Not "not ever." Just "not yet."

It's not the advice they expect. Most people come to us because they've decided AI is something they should be doing. They've read the articles, seen the competitors, felt the pressure. When we say "hold off for now," there's usually a pause followed by "really?"

Yes, really. And here's why that's actually good news.

The pressure to "do AI" is real

There's a narrative in business media right now that if you're not using AI, you're falling behind. Every conference, every trade magazine, every LinkedIn post seems to carry the same message: adopt AI or get left behind.

Some of that is true. AI is genuinely useful for many businesses. But the urgency is often overblown, and it leads to businesses rushing into projects they're not ready for.

The OECD's SME and Entrepreneurship Outlook makes an important distinction between AI readiness and AI urgency. Many businesses would benefit from AI eventually, but not all of them are in a position to benefit from it right now.

When we say "not yet"

Here are the most common reasons we advise a business to wait:

Your processes aren't defined

AI automates processes. If your processes aren't clear, consistent, and documented (even informally), there's nothing solid to automate. You'll end up building AI on top of chaos, which just produces faster chaos.

This is the most common reason we pump the brakes. A business comes to us wanting to automate their order processing, but when we look at how orders actually get handled, it's different every time. Different people do it differently. There are workarounds built on workarounds. Half the knowledge is in one person's head.

The fix isn't AI. The fix is process improvement. Get the process right first, then automate it. We can help with both, but they need to happen in the right order.

Your data is a mess

AI needs data to work with. If your customer records are incomplete, your product catalogue is out of date, or your financial data lives in six different spreadsheets that don't agree with each other, AI will struggle.

This doesn't mean your data needs to be perfect. But it does need to be basically reliable. If your team doesn't trust the data in your current systems, AI won't magically fix that. As the UK's National Data Strategy emphasises, data quality is the foundation for any data-driven initiative, AI or otherwise.

You're in the middle of something else

Businesses have limited capacity for change. If you're in the middle of a system migration, a restructuring, a major product launch, or any other project that's consuming management attention, adding an AI project on top is a bad idea.

AI projects need engagement from the team, attention from leadership, and bandwidth for testing and feedback. If those things aren't available, the project will stall or produce poor results.

The ROI isn't there yet

Sometimes we look at a business and the honest answer is: the potential savings don't justify the investment right now. If you're a 10-person business with relatively low volumes, automating a process that takes two hours a week might cost more to implement than it saves in the first year.

That doesn't mean it won't make sense later, when the business is bigger, or when the technology gets cheaper (which it is, rapidly). But right now, your money is better spent elsewhere.

Your team isn't ready

This one is sensitive, but it matters. If your team is already stressed, already resistant to change, or already dealing with recent upheaval, pushing AI on them will backfire. People need to be in a reasonably stable place before they can absorb new ways of working.

The CIPD's guidance on organisational change is clear: change fatigue is real, and stacking changes on top of each other leads to disengagement and resistance.

What we recommend instead

When we tell a business to hold off on AI, we don't just leave them hanging. We tell them what to do in the meantime so they're ready when the time is right.

Tidy up your processes

Document how things actually work today. Not the ideal version. The real version. Then look for obvious improvements, steps that can be eliminated, handoffs that can be simplified, exceptions that can be standardised.

This work has immediate value even without AI. And when you do eventually automate, you'll be automating a clean process rather than a messy one.

Sort out your data

Pick your most important system (usually your CRM or your accounting software) and get the data in order. Remove duplicates, fill in gaps, establish who's responsible for keeping it accurate. Again, this has value on its own.

Build AI literacy

You don't need your team to understand machine learning. But it helps if they understand what AI can and can't do in practical terms. Encourage people to experiment with tools like ChatGPT for low-risk tasks: drafting emails, summarising documents, brainstorming ideas. This builds familiarity and reduces fear.

Keep a list

As your team spots repetitive, time-consuming tasks, write them down. This becomes your automation shortlist for when you're ready. The best ideas often come from the people doing the work.

Why this is good news

Being told "not yet" feels like a setback. It isn't. Here's why.

A business that takes six months to get its processes and data in order, and then starts an AI project, will get better results faster than a business that dives in unprepared. We've seen it repeatedly. The prepared business gets a clean implementation, quick adoption, and measurable ROI. The unprepared business gets delays, frustrations, and a team that associates AI with failed projects.

There's also a cost argument. AI tools are getting cheaper and more capable every quarter. A project that costs £15,000 today might cost £10,000 in twelve months and deliver better results. If you're not ready today, waiting doesn't just buy you preparation time. It often buys you better technology at a lower price.

How to know when you are ready

Here's a simple checklist:

• Your key processes are consistent (different people do them roughly the same way)
• Your core data (customers, products, financials) is reasonably accurate and up to date
• Your team has capacity for some change without being overwhelmed
• You can identify at least one repetitive, time-consuming task that's clearly suited to automation
• Leadership is prepared to give the project attention and support

If you can tick all five, you're ready. If you can tick three or four, you're close, and a short preparation phase will get you there.

Our commitment

We'd rather tell you to wait and come back in six months than take your money for a project that's going to struggle. That might sound unusual for a consultancy, but it's how we work. Our reputation depends on projects that succeed, not projects that start.

When you are ready, we'll show you exactly where to start. We handle the technical side entirely. You bring a business that's prepared for improvement. We bring the tools and experience to deliver it.

Get your free AI opportunity report and we'll give you an honest assessment of where you stand, including what to work on now so you're ready for AI when the timing is right.