For years, being good at this job looked like a particular set of skills. You knew the status before anyone asked. You could chase five people for updates without making enemies. You could turn a messy fortnight into a clean two-page report by Friday. Those are exactly the tasks AI tools are now being pointed at, and if they were how you knew you were good, watching them get automated does not feel like a productivity gain. It feels like someone deleting the part of the job you were proud of. This post takes that feeling seriously rather than trying to talk you out of it.
It is not a fringe reaction. The trust gap is measurable: in Smartsheet's 2026 survey of 1,651 project and portfolio professionals, 97% were experimenting with AI, and only 46% trusted it to act without human supervision. And some firms are going further than scepticism:
"...some companies beginning to position themselves as 'AI-free' as a mark of quality..." James Garner, APM, 20 January 2026
Two disclosures before the argument. Smartsheet sells project management software. And we sell a course on building AI systems for Scrum Masters and project managers, so read what follows knowing which way our interests point.
What this post can and cannot prove
The numbers above show that trust is low and that some resistance is organised. They do not show why any individual resists, and we have found no study showing that resistance is strongest among the most experienced people, or that it comes from professional identity. That part is our argument, not a finding. It is built from what we hear and from what the work itself looks like. Judge it on whether it describes you.
The argument is this. Much of what made a project professional visibly good was assembly: collecting status, chasing updates, formatting the report, keeping the register current. Assembly is the part that AI tools handle best, because it is the part where two competent people would produce the same output. But assembly was never the whole of it. It was the visible part of something else.
Done honestly, the exercise can be a surprise in either direction. The right column may be longer than you expected. The left column may turn out to be the part you quietly disliked.
"If AI writes the status report, what am I actually for?"
I have spent ten years being the person who knows where everything stands. If a tool knows that now, I am not sure what is left.
Look at your inventory's right column. A tool that writes the status report knows what the tickets say. It does not know that the green item is green because someone stopped asking difficult questions, or that the sponsor needs the bad news on Tuesday rather than in Friday's report. Knowing where everything stands was never really about the collecting. It was about what you did with it.
When the collecting is handled, the part that remains is the part people actually valued, now without the hours of assembly around it.
"Firms are marketing themselves as AI-free. Maybe resisting is right."
If serious firms are selling "no AI" as a quality mark, then refusing it is not stubbornness. It is professional judgment.
It can be. The APM piece is right that the positioning exists, and the trust numbers say most professionals share the doubt. Refusing to let an unchecked tool write anything that reaches a client is a defensible standard.
But a firm's market positioning and your own working method are different questions. You can hold the AI-free standard for anything that leaves the building and still hand the assembly to a tool, with checks, for the work that feeds your judgment. Scepticism about what a tool produces is exactly the attitude that makes someone good at checking it. Our post on AI pilots describes a simple test for which tasks are safe to hand over.
"I have seen hype cycles before. This one will pass too."
Every few years something is going to change project management forever. Most of them did not. I would rather wait this one out.
That has been good judgment before, and it may be again. Nobody, including us, can tell you how much of today's enthusiasm survives. What is different this time is the scale of the trial: when nearly everyone in the field is already experimenting, the question is less whether you will meet these tools and more on whose terms.
The inventory is useful whichever way it goes. If the cycle passes, you have lost an hour and gained a clearer description of your own value. If it does not, you have it before you need it.
The part worth keeping
The fear is reasonable. If the visible part of your work was assembly, automation will feel like it is taking the craft. But the craft was always the judgment the assembly served: knowing what it means, who needs it, and how to say it. That part is still yours, and it may well be the bigger part, once it is written down where you can see it.
Sources: Smartsheet, 2026 Project and Portfolio Management Priorities Report | APM, Five AI trends for 2026 that project managers need to consider. Checked 2 October 2026.
