Evidence current to mid-2026.
In project management, AI is ready today for the documentation overhead, drafting status reports, capturing meeting actions and maintaining the Risks, Assumptions, Issues and Dependencies (RAID) log, always with the project manager (PM) reviewing before anything goes out. Estimates, critical-path calls and stakeholder politics stay with people. The tables below show exactly which is which.
Each task below sits in one of three bands: a strong fit today, worth a careful pilot, or keep with people for now.
A clear breakdown of where AI does and does not fit across the core tasks of a project manager. A practical starting point, not the last word.
Proven and available today. AI does the bulk of the work and the project manager reviews it.
| Function | The job today | With AI |
|---|---|---|
| Status report drafting | Pull updates from the team, write the weekly status narrative, send to the steering committee | AI drafts the status report from notes, tickets and prior reports; the project manager checks the facts and sends |
| Meeting notes and action capture | Listen, write notes, pull out actions with owners and due dates, circulate after the meeting | AI transcribes and extracts actions, owners and decisions; the project manager checks the list is correct and complete |
| Risks, Assumptions, Issues and Dependencies (RAID) log drafting and maintenance | Capture risks, assumptions, issues and dependencies from team input and meetings, write them up, keep the log current | AI drafts RAID entries from meeting notes; the project manager owns the ratings, treatment decisions and keeps the log |
| Stakeholder updates and comms | Write the same update in different registers for the steering committee, the sponsor, the team and the client | AI drafts each version from the source narrative; the project manager adjusts tone, checks accuracy and sends |
Promising but not yet proven at this scale. AI assists and the project manager stays in the loop, so trial it on a contained scope first.
| Function | The job today | With AI |
|---|---|---|
| Project plan and Work Breakdown Structure (WBS) first drafts | Build the work breakdown structure from scratch, sequence tasks, assign owners, set durations | AI generates a first-draft task list and WBS from a brief; the project manager restructures, adds context and owns the sequencing |
| Schedule narrative and re-plan summaries | Explain what moved, why, and what the new plan means for the delivery date | AI drafts the re-plan summary from before/after schedule data; the project manager adds the real drivers and owns the stakeholder message |
| Business case and project brief drafting | Structure and write the case for investment: problem, options, costs, benefits, risks, recommendation | AI drafts the skeleton and standard sections from a brief; the project manager owns the numbers, the options and the recommendation |
| Lessons-learned synthesis | Gather retrospective notes and post-project reviews, find the patterns, write the summary | AI synthesises themes across multiple inputs; the project manager validates and owns the organisational recommendation |
| Dependency and milestone tracking | Keep a live register of what this project needs from others and what others need from it; chase for updates | AI drafts status summaries and flags missed milestones from data; the project manager investigates and owns the recovery conversation |
These stay with people, either because the judgement required is irreducibly human or because accountability cannot be delegated.
Worth weighing: for the strong-fit tasks, the upfront setup (templates, building the assistant’s context, connecting to your project data) is a one-off. The ongoing effort on those tasks is mostly review, not the old grind of writing from scratch each week. The time that comes back goes to the parts AI cannot touch: reading the room in a difficult stakeholder conversation, making the judgement call on what to cut when the schedule slips, and being present in the delivery rather than catching up on documentation.
One honest gap worth naming: an AI-drafted status report can look polished and confident even when the underlying project is in trouble. The draft is only as good as the input you give it, and a thin or over-optimistic input produces a thin or over-optimistic report. The project manager still owns the forecast, the risk assessment and the escalation call. The AI handles the formatting, not the judgement.
The evidence
The research, regulators and tools behind this guide:
This guide sits on top of the things that stay the same whatever your function. See the ground rules
Questions
No. The parts that take the most time, writing status reports, typing up meeting notes, maintaining the RAID log, are real candidates for AI assistance. The parts that make a project manager valuable, reading the room, making the hard call on what to cut, owning the delivery commitment, are not. What shifts is the ratio of documentation to delivery. Less time writing, more time leading.
It depends on the data and the client. For internal drafting where a PM reviews before anything leaves, the risk is low. For recording external meetings, you should obtain consent from every participant before the recording starts; OAIC guidance frames this as best practice and it is the right default for external stakeholders and clients. For government or regulated projects, check the agency's AI use policy and your data classification before using any cloud tool. The rule of thumb: use AI to draft, never to send without review, and never paste client names or contract details into a public tool.
Meeting notes and action capture is the most practical first step. Set up a transcription tool on your internal team meetings (Teams, Zoom, or a standalone tool), review the action list for the first week, and build confidence in what it gets right and wrong. Status report drafting is the next obvious step once you have clean meeting notes to feed it.
A general assistant can produce a useful first-draft work breakdown structure from a brief, which is a better starting point than a blank page. The sequencing, durations and commitments remain human work. No AI tool can know this team, this client or this environment well enough to own the plan. Use it to start, not to finish.
Yes, the tasks in the strong-fit and pilot bands apply across all three. The documentation and coordination overhead is the same whether you are delivering a system, a fit-out or an organisational change. The WHS seam applies on physical projects: safety content goes to the WHS function, not the project management toolkit.
The evidence is current to mid-2026. Asana and Monday.com AI features reached general availability in the first half of 2026. Microsoft 365 Copilot in Teams for meeting notes and task management is broadly available on eligible plans. Other tools, including Linear, are in active beta with some features locked to higher tiers. This is a fast-moving area. Where we use a number we say where it comes from; vendor time-saving claims are treated as upper bounds until independent evidence says otherwise. If something has shifted, the fastest way to get the current picture for your context is a quick chat.
Book a discovery callA first conversation
Book a confidential chat. Tell us where you see the opportunity, in project management or anywhere else, and we will agree a sensible next step. No deck, no pitch.
Closest to project management
See all guides