A way to bring in AI and new process without it feeling like an erasure of what got the business here.
Key takeaways
For a lot of family businesses, "the way we do things" isn't just habit, it's the accumulated result of decisions that built the business's reputation, often made by someone still in the room. Proposing to change it can land less like an operational suggestion and more like a judgment on everything that came before.
That reaction is worth taking seriously rather than working around. A modernisation effort that ignores it tends to meet quiet resistance everywhere, even where the change itself makes sense.
The framing that tends to work is additive: AI and new process sit on top of the business as it already runs, handling the parts that were always tedious rather than replacing the judgment that made the business what it is. Documenting a pricing decision doesn't remove the owner's discretion, it just means someone else can see how that discretion has been applied.
This distinction isn't just semantics. A tool introduced as "this replaces how you used to do it" gets resisted; the same tool introduced as "this handles the paperwork so you can keep doing the part that matters" tends to get adopted.
The person whose knowledge is being captured should be shaping how it's captured, not having a system built around them without their input. Involving them in deciding what gets documented first, and in what order, turns the process from something being done to the business into something the owner is actively leading.
This also produces better documentation. An owner who feels ownership over the process tends to give fuller, more honest answers than one who feels like they're being audited.
The instinct to start with the business's most defining process, the thing everyone agrees is "how we built our reputation", is usually the wrong one. That process carries the most emotional weight and the least tolerance for early mistakes. A better first project is whatever is currently the most painful and lowest-risk: a scheduling headache, a reporting task nobody enjoys.
An early, visible win on something nobody's attached to builds the trust needed to eventually touch the processes that actually matter most to how the business sees itself.
A 30-minute call is enough to tell you whether AI pays for itself here.