Perspective · Adoption
Getting frontline buy-in for AI.
You can build the best AI tool in the world and watch it die on the floor because the people who were supposed to use it never bought in. In my experience, AI adoption almost never fails on the technology. It fails on people — and that's both the bad news and the good news, because people problems are solvable if you treat them as the main event, not an afterthought.
Why the frontline resists (and why they're often right)
When frontline staff are cool on a new AI tool, it's rarely stubbornness. Usually they're reacting to something real: a fear that it's there to replace them, a tool that was designed without asking how the work actually happens, or memories of the last "transformation" that made their day harder. If you dismiss that as resistance, you lose. If you treat it as feedback, you usually find they're pointing at a genuine flaw in the rollout.
Involve them before you build, not after
The single biggest predictor of adoption is whether the people who do the work were involved early. The team that handles the exceptions knows where the time actually goes, which cases are weird, and what "good" looks like — knowledge no consultant or executive has. Bring them in during discovery, not at training. People support what they helped shape, and they'll quietly sabotage what's done to them.
Frame it as help, and mean it
"This will make you more productive" sounds, to a frontline ear, a lot like "we'll need fewer of you." The framing that works is honest and specific: this takes the tedious part off your plate so you can do the part that needs a human. And it has to be true — the best AI rollouts genuinely remove the grind (the chasing, the re-keying, the boring triage) and leave people the judgment, the relationships, and the interesting work. If your plan is really about cutting heads, the team will know, and no amount of framing will save it.
Make the first win theirs
Pick a first project that solves a problem the team actually complains about — not the one that looks best to leadership. When the tool kills a chore everyone hates, you don't have to sell adoption; people pull it toward themselves and tell their colleagues. Let a respected team member be the early champion, give them credit for the win, and the rest of the floor follows far faster than any mandate would move them.
Keep humans in the loop — visibly
Trust is built when people can see that the AI drafts and suggests, but they decide. Designing clear human review into the workflow isn't just good governance; it's good change management. It tells the team, concretely, that they're still in control — which is exactly what they're worried about. The "human in the loop" you build for safety doubles as the reassurance that earns buy-in.
The unglamorous truth
The technology is the easy 20%. The people side — involving them, framing it honestly, choosing a first win they care about, and keeping them in control — is the 80% that decides whether your AI investment lives or dies. Budget time and attention for it accordingly, and most of the "AI doesn't work here" stories simply don't happen to you.
Want adoption built in from day one?
An AI Opportunity Assessment doesn't just find the workflow — it plans the owner, the first win, the human-review design, and the rollout that gets your team on board. About two weeks.