Some clients are not confused about AI, they are skeptical of it, often for good reason. Here's how I earn trust with clients who start the relationship doubting the whole premise.
A client once told me, in our first call, that the last “AI expert” she hired had cost her three weeks and a broken workflow, and that she was only taking my call because a colleague had insisted. That is not skepticism you overcome with a good pitch. It is skepticism you overcome by being right, repeatedly, in small verifiable ways before you ever ask for anything larger.
Most advice about selling AI services assumes the client just needs clearer explanations. Some clients do. But a growing number are skeptical for a completely rational reason: they have already been burned, either by a vendor who overpromised or by their own attempt at using off-the-shelf AI tools that produced confidently wrong output. Explaining the technology better does not fix that. Demonstrating restraint does.
The instinct when facing a skeptical client is to explain harder. More detail about the architecture, more reassurance about how careful you are. This mostly backfires, because a client who has been burned before is not missing information. They are pattern matching against a bad prior experience, and pattern matching does not update on explanation. It updates on evidence.
I stopped trying to talk skeptical clients out of their skepticism. Instead, I agree with the parts of it that are correct. Most AI implementations do overpromise. Most vendors do underestimate the maintenance burden. Saying that out loud, unprompted, does more for trust in the first ten minutes than any amount of technical credibility signaling.
The exact architecture, memory layers, and delegation patterns I use to run 50 agents across two businesses.
Get the AI Agent Blueprint →Start smaller than they expect, on purpose. With skeptical clients, I deliberately propose a narrower first engagement than what I think they eventually need. A single automated task, low risk, easy to verify. Not because it is the ideal scope, but because a small, verifiable win rebuilds trust faster than a large, ambitious pitch ever could. The larger engagement gets proposed after, once there is evidence to point to.
Show the failure modes before they ask. I walk skeptical clients through exactly how the system could go wrong and what happens when it does, before they ask. Where the human review checkpoints sit. What the dead letter queue is and how often something lands in it. Clients who have been burned before are specifically listening for whether you are hiding the failure modes or being upfront about them, and that single signal often matters more than anything about capability.
Give them a kill switch and mean it. Every engagement includes an explicit, easy way to pause or shut off the automation entirely, with no argument and no penalty. I have had exactly one client use this, briefly, during a period of internal reorganization. The fact that it exists and works as described mattered more to every other skeptical client than the feature itself ever got used.
Report on what did not work, not just what did. My monthly updates to skeptical clients include a section on edge cases the system flagged for human review, near misses, and anything that needed a manual fix. Clients who trust AI easily rarely ask for this. Clients who do not trust it read this section closely, and it is usually what convinces them the relationship is honest.
Clients who start skeptical and get won over through evidence become the steadiest long-term relationships I have. They already did the work of scrutinizing the claim before agreeing to anything, so once they are in, they are in with clear eyes rather than enthusiasm that can deflate at the first hiccup.
You do not overcome real skepticism with a better pitch. You overcome it by being the vendor who was honest about the risk before anything went wrong, and then correct when something eventually did.
The newsletter covers real client engagement stories, including the ones that started with a hard no.
Subscribe to The Signal
One issue per week. The biggest AI build of the week, the top five videos worth your time, and an operator's read on what actually matters. For builders who ship, not dreamers who scroll.
No spam. Unsubscribe anytime. Just The Signal.