Talk to enough business owners about AI and a pattern emerges: the tools aren’t the problem. The scoping is.
The trap of starting too broad
“We want to use AI across the business” is not a project — it’s a mission statement. Without a specific, painful, well-understood process to target first, AI initiatives tend to sprawl, lose momentum, and quietly die in a Slack channel nobody checks anymore.
What actually works
The businesses that get real value start narrow: one process, one team, one measurable outcome. Get that working, prove the time saved, then expand. It’s unglamorous, but it’s the difference between a pilot that sticks and a pilot that becomes a cautionary tale in next year’s budget meeting.
Human driven, not human replaced
The other stall point is trust. Teams disengage from automation that feels like it’s replacing them rather than supporting them. The projects that succeed treat AI as leverage for the humans already doing the work — removing the repetitive parts of their job, not their job itself.
That’s the whole idea behind “AI powered, human driven.” The technology does the repetitive heavy lifting. The judgement, relationships, and decisions stay with your team.