AI and Your People in a Family Firm

A lot of Midlands manufacturing sits in businesses that are second or third generation, where people have been there twenty or thirty years and the knowledge of how things work lives in their heads.

That is simultaneously the firm's greatest asset and the hardest thing about introducing AI — and treating it as a change management exercise borrowed from a corporate playbook will not work.

What people actually fear

Not usually redundancy, in a firm where they know everyone. More often:

That their judgement will be overruled by a machine — a real concern for someone whose standing rests on being the person who knows.

That thirty years of experience is being quietly extracted and made replaceable.

That they will look incompetent in front of younger colleagues while learning something new. This is the one nobody says out loud, and it is often the strongest.

None of those are addressed by explaining the technology better.

What works

Make the expert the author, not the subject. If your senior estimator helps build the system — deciding what it should consider, correcting its drafts — it becomes a record of their judgement rather than a replacement for it. Same tool, entirely different meaning.

Start where the work is disliked. There is always a task everyone hates: goods-in paperwork, chasing delivery confirmations, retyping the same specification. Automating a disliked task makes the technology an ally on day one.

Be exact about jobs. Vagueness gets filled with the worst interpretation. If no roles are going, say it plainly. If something is changing, say what, and say it early — people forgive change far more readily than they forgive finding out late.

Let people see it be wrong. Counter-intuitive, but watching the system produce a bad answer and be corrected does more for trust than any demonstration of it working. It converts a mysterious oracle into a tool with limits.

What backfires

Bringing in a consultant who talks about transformation to people who have run the same machines for two decades. Announcing it at an all-hands before speaking to the people affected. Piloting it with the youngest member of staff and presenting the result to the most experienced. And describing it as removing human error to a room of humans.

The succession angle

There is a version of this conversation that lands well in a family firm, and it is worth having openly.

If the person who knows how to price the difficult jobs retires in three years, that knowledge leaves with them. Capturing it — in a system that drafts, explains its reasoning and can be corrected — is succession planning. It is the same activity as writing it down, except it actually gets done, because the system is useful to them now.

Framed that way, the experienced people are not the obstacle. They are the reason to do it, and the only ones who can.

A reasonable pace

One tool, one team, visible results, then the next. Firms that go slowly and take people with them end up further ahead in two years than firms that announce a programme, because they are not spending the second year rebuilding trust.