AI Competency Framework for Project Managers: A Starting Point
An AI competency framework for project managers maps AI capability across the disciplines PMs already own: prioritisation, stakeholder communication, risk management, delivery tracking, at novice, intermediate, and advanced levels, so “AI-literate” means something concrete instead of whatever the interviewer happens to have in mind. Novice use looks like treating AI as a faster search box. Advanced use looks like designing AI-augmented workflows that change how the team operates, not just how fast you personally work.
If you’ve ever been asked how you use AI day to day, given a perfectly reasonable answer, and felt it didn’t quite land, this might be why.
”AI-Literate” Means Nothing Right Now
Picture the interview. A PM is asked how they use AI. They talk about using ChatGPT for meeting notes, drafting emails faster. A reasonable answer. They don’t get the role, and never find out why.
That answer might have been fine. It might also have read as exactly the novice-level use described below, depending entirely on what the interviewer had in mind. Neither the candidate nor the interviewer could tell you which, because there’s no shared vocabulary for what “good” looks like at each level.
“AI literacy” has become a screening criterion without becoming a defined one. That’s a problem for candidates assessed against an invisible bar, and for PM leaders trying to coach a team toward it.
An AI Competency Framework Across the Disciplines You Already Own
This maps onto the PM disciplines that already exist, not a new set of categories: prioritisation, stakeholder communication, risk management, delivery tracking. Within each, there’s a meaningful difference between novice, intermediate, and advanced AI use.
Prioritisation. Novice: asking AI to help prioritise the backlog and taking the output mostly as is. Intermediate: feeding AI structured context (effort, dependencies, stakeholder weight) and treating its output as one input into your own judgement. Advanced: a repeatable AI-assisted prioritisation workflow the whole team uses, with the reasoning visible and checkable, not just the ranked list.
Stakeholder communication. Novice: AI drafts an update, you tidy the wording. Intermediate: AI drafts tailored to specific stakeholder context you’ve supplied, what they care about, what’s changed since last time. Advanced: stakeholder updates generated consistently from project data with minimal manual assembly, freeing your time for the conversations the updates are actually about.
Risk management. Novice: asking AI what the risks are on this project and getting generic categories back. Intermediate: feeding AI your actual risk register and recent project context for synthesis and gap-spotting. Advanced: a trust-calibrated workflow, the kind covered in an earlier piece in this series, where AI-flagged risks are verified against sources before they reach a steering committee.
Delivery tracking. Novice: asking AI to summarise status from whatever’s pasted in. Intermediate: AI synthesising status from multiple structured sources, standups, tickets, commits. Advanced: AI cost and capacity tracking built into how the team monitors delivery, not bolted on separately.
The Real Line: Personal Speed vs. Team-Level Change
Notice the pattern across all four. The novice-to-intermediate jump is mostly about input quality: better context in, better output out. That’s real progress, and reaching it consistently across all four disciplines is a genuinely useful bar to clear.
The intermediate-to-advanced jump is different in kind, not just degree. It’s the move from “I do my own work faster” to “the way the team operates has changed because of how AI is built into the workflow.” Reaching it means designing a workflow for other people, not just optimising your own.
That reframes what “AI-capable” actually screens for: whether you think about your work as a system you can redesign, which has always been a core PM skill and which AI just makes visible faster.
Score Yourself Across the Four Disciplines
For example:
Discipline: Prioritisation Current level: Intermediate, feeds AI structured backlog context, still makes the final call solo Target level: Advanced, a workflow the team runs together Next step: Write down the process already in use and hand it to the team as a starting workflow
Discipline: Stakeholder communication Current level: Advanced, updates generate consistently from project data with minimal manual assembly Target level: Already there Next step: None needed right now
Discipline: Risk management Current level: Novice, asks a bare “what are the risks” question and gets generic categories back Target level: Intermediate, feed it the actual risk register and recent context instead Next step: Start feeding the real register in rather than asking the bare question
Discipline: Delivery tracking Current level: Intermediate, AI synthesises status from a few structured sources Target level: Advanced, integrated cost and capacity tracking Next step: Lower priority for now than fixing risk management first
Now score your own four disciplines the same way:
Discipline: [prioritisation / stakeholder communication / risk management / delivery tracking] Current level: [novice / intermediate / advanced, with one line on why] Target level: [where you actually want to be] Next step: [the one concrete thing that gets you there]
Three Uses, One Scorecard
This same scorecard works in more than one direction. Coaching a team, it gives you a shared vocabulary for development conversations that currently happen on vibes: “you’re intermediate on stakeholder comms, what would advanced look like for your specific projects” beats “get better at AI.” Hiring, it gives you something concrete to ask about instead of a generic “do you use AI,” calibrated against real levels instead of how sophisticated the answer sounds.
What this actually does is give how PMs work with AI a vocabulary precise enough to talk about, rather than leaving it as a vague vibe check nobody can act on.
That clarity, knowing exactly where you sit and what the next concrete step looks like, matters whether you’re novice or already advanced.
Frequently Asked Questions
What does “AI-literate” actually mean for a project manager? There’s no single industry-standard definition, which is part of the problem. A useful framework maps it across core PM disciplines (prioritisation, stakeholder communication, risk management, delivery tracking) at novice, intermediate, and advanced levels, where advanced means designing AI-augmented workflows for the team, not just working faster yourself.
What’s the difference between intermediate and advanced AI use for PMs? Intermediate is about input quality, feeding AI more structured context to get better personal output. Advanced is a different kind of change: redesigning how the team’s workflow operates around AI, not just optimising your own work.
How can I use an AI competency framework if I’m hiring project managers? Ask concrete questions about specific disciplines, such as how you’d use AI in stakeholder communication, rather than a generic “do you use AI” question, and calibrate against novice, intermediate, and advanced examples.
Where would you place yourself across those four areas, and is it the same level everywhere, or does it vary more than you’d expect?
Yes, AI helped me to write this :)