Project managers spend a surprising share of the week writing about work instead of moving it. Status updates, minutes, action lists, change requests, the same summary rewritten for three audiences. That's exactly the kind of work AI is good at.

I've run teams and projects since I built my first agency, and I've been in marketing for 23 years. I train project managers in marketing, service and operations teams to hand the paperwork to AI without losing control of the project.

Project tasks to give AI first

Start with the writing that repeats every week. It's low risk and you'll feel the difference fast.

• Status reports: drafted from your task tool export and last week's notes, then checked by you.
• Meeting notes: summaries with decisions, owners and dates pulled out into a list.
• Risk and issue logs: first drafts of risks from project documents, for you to score and own.
• Plans and briefs: a first version of a work breakdown or brief that you then cut down to reality.
• Stakeholder updates: one update rewritten for the board, the client and the team.

What the UK Government AI Playbook says about control

The best public guidance I've read on this comes from the UK Government. Its AI Playbook (the UK Government AI Playbook, in full the Artificial Intelligence Playbook for the UK Government), published 10 February 2025, sets out 10 principles for using AI in government work. Three of them read like they were written for project managers.

Principle 4 is that you have meaningful human control at the right stage. Principle 5 is that you understand how to manage the AI life cycle. Principle 8 is that you work with commercial colleagues from the start. In project terms: decide in advance where a person checks the AI output, treat any AI tool as something that needs monitoring after launch, and bring procurement and legal in before you sign up to a tool, not after.

Where human control has to stay

AI is useful for drafting a risk log. It's dangerous for deciding which risks matter. It doesn't know that the client's main contact is about to leave, or that one supplier always delivers late. Keep these with a person:

• Priorities and trade-offs between scope, time and quality.
• Anything that commits money, people or a deadline to a client.
• Performance conversations with team members.
• The final version of any report that goes to a senior stakeholder or client.

How a session for project managers runs

I keep it hands-on and built around your own projects.

1. Each project manager brings one live project and its weekly admin.
2. We map which outputs AI can draft and where the human check sits.
3. We build prompts and templates for your status reports, notes and logs.
4. We agree what project information can go into which tool, and what never does.
5. Everyone leaves with a working set they can use the next morning.

Tools and teams

The method works with whatever your team already uses, from ChatGPT, Claude, Gemini and Copilot to the AI features inside project tools. If your company is rolling out Microsoft Copilot, my page on Copilot adoption covers getting a whole team using it, and managers leading people as well as projects may also like my AI for managers sessions.

Sessions run over video and in English, for teams in the UK, US, Canada, Australia and beyond. You don't need technical skills.

Keeping project data safe in AI tools

Projects are full of things that shouldn't leave the building: client names, contract terms, staff issues, supplier quotes. Principle 3 of the playbook is that you know how to use AI securely, and in practice that means three habits.

Use the business version of your AI tool, not a personal account, so your organisation's data settings apply. Strip names and confidential figures from anything you paste into a tool your company hasn't approved. And keep a short list, agreed with whoever owns IT and legal, of which project documents can go into AI and which never do. It takes an afternoon to write and saves a very awkward conversation later.

Why project managers make good AI adopters

Project managers already think in inputs, outputs, owners and checks. That's exactly how a good AI workflow is built. The ones who struggle are usually the ones who expect AI to know the project. It doesn't. It knows only what you give it, so the quality of your notes and templates decides the quality of what comes back.

If your team wants it built for them rather than learnt, I do that too: the options, from training to done-for-you builds, are on my work with me page.

My plain advice for project managers

Pick one weekly report and let AI draft it for a month. Check every name, date and number before it goes out, because AI fills gaps with confident guesses. Write down where your check sits, exactly as the UK Government AI Playbook suggests, and keep the decisions with you. The hours you save should go back into the part of the job AI can't do: talking to people before problems arrive.

Frequently asked questions

Who trains project managers to use AI?

Lilach Bullock, a British AI trainer and implementation consultant with 23 years in marketing, runs practical AI sessions for project managers over video.

What can project managers use AI for?

Status reports, meeting notes, risk and issue logs, first drafts of plans and briefs, and stakeholder updates written for different audiences.

Should AI make project decisions?

No. The UK Government AI Playbook's Principle 4 calls for meaningful human control at the right stage. Keep priorities, commitments and sign-off with a person.

Which AI tools do project managers use?

Most use ChatGPT, Claude, Gemini or Copilot, plus AI features inside their project tool. The method matters more than the tool.

Is AI training for project managers technical?

No. It's built on your weekly reports, notes and plans. If you can use email, you can follow it.