We have talked to more than fifty small and mid-sized businesses about AI this year. The same three things come up almost every time, and none of them are about the technology.
1. "Nobody has told us how we should be using it"
We hear this most from people in the middle of the business rather than at the top. They know AI exists and plenty have tried it. What they have never been given is any steer from their own company on where it fits.
Leadership is usually surprised by this. They have discussed it, maybe approved some licences, and assume everyone got the message. Most of the time it has not reached them. Buying licences is not the same as telling people what to do with them.
2. "We are worried about getting it wrong"
It is not the technology people are worried about. It is looking foolish. Sending a client something with a mistake in it that nobody picked up. Being seen to cut corners on work they are paid to do properly.
There is another one that comes up less often, usually right at the end of a meeting. If I get really good at this, am I training my replacement? People want a straight answer to that, not reassurance.
3. "How safe is our information?"
Nobody asks this in technical terms. It usually comes out as "I do not really know what happens to what I type into it."
On a business plan the answer is a good one. Your data stays in your own tenant and the major providers do not use it to train their models. On a personal account the terms are not the same. Most staff have never had that explained to them.
THE BIGGEST CONCERN
Where there was no direction from the business and no company tools provided, staff were using their own personal AI to do company work
They were not trying to cause a problem. They had deadlines, found something that made their life easier, and nobody had told them not to.
But company information is going into accounts the business has no agreement with and no visibility of. Client names, contract terms, pricing, staff data. If a client or a regulator asks what was shared, you have no way of answering. Organisations need to address this, and quickly.
What to do about it
Ask what people are already using. Make it clear nobody is in trouble. The answer is usually more than you expect, and there is no point planning from a starting position you have got wrong.
Name two or three tasks. Specific, low risk, and in roles where the time saved shows up. Saying you want to be an AI-enabled business gives nobody anything to do on Monday morning.
Write the boundaries down. One page is enough. What can go in, what cannot, which accounts to use and who to ask. That deals with most of the fear in the second point and most of the risk in the third.
You do not need a consultant for any of that. The businesses getting real value out of AI are the ones that decided what they wanted it for.
