AI in the PMO: a three-hour working session for your team
Your people are already using AI, and the PMO’s reporting and governance probably look much as they did a year ago.
This is a remote, hands-on session for a whole PMO or delivery team. In three hours you sort your own recurring work by where AI belongs, take one of your own processes apart together, and each person names an experiment to run. The price is £1,999 + VAT for the whole team.
What the session is
Most PMOs we talk to are in the same place. Individuals use Copilot or a similar tool for meeting notes and first drafts, and feel faster for it. The PMO’s reporting, governance and follow-through look much as they did a year ago. The Work AI Index 2026 put numbers on that gap: of the 6,000 digital workers surveyed, 87% use AI at work and 13% said their organisation performs “significantly better” as a result.
The session is built for a whole team in one sitting, from people who have never opened Copilot to people already building agents. The material comes from our July 2026 webinar, AI for PMOs, and from running the session in September 2026 for a portfolio office of more than 20 project managers, delivery managers, business analysts and change managers. In that room, eight of the 20 people who answered the opening poll had used Copilot rarely or never, and three were already building things. The design assumes that spread.
Every exercise uses the team’s own tasks, reports and processes, and the AI tools your organisation has already approved.
Who it is for
The session suits a Head of PMO or delivery lead who recognises one of these.
A team of mixed ability
Some of your people use AI daily and some have never tried it. Anything pitched at one end loses the other, so the session works on shared tasks and lets each person start where they are.
Individual speed, no team gain
People say they are faster. The status pack, the RAID log and the steering committee papers take as long as they ever did and vary from month to month.
Tools approved, rules unclear
Copilot or a similar tool is available and permitted, and nobody has agreed what may go into it, what must be checked, or which tasks it should be kept away from.
What happens in the three hours
Timings flex with the room, and the blocks below are designed so that each one leaves the team with something written down.
Where AI belongs, and where it does not
The survey evidence on individual speed against organisational results, then the two questions that decide where a task sits: how costly is an error, and does the task run on explicit information or tacit judgement? Worked PMO examples in each of the four zones. The framework comes from Anand and Wu’s “The Gen AI Playbook for Organizations”, Harvard Business Review, November 2025.
Working safely and briefing the tool well
Your organisation’s tool policy, what must never go into a prompt, and Microsoft’s four-part structure for a brief: goal, context, expectations and source. We then build a status update in front of the room from a realistic set of source material (two emails, a risk log, a change request and a set of RAG definitions), and the room judges whether the draft is one it would send to a director.
Your team's week, sorted
Everyone adds their own recurring tasks to a shared online board and places each one in a zone. We read the board together and argue about the ones near the lines. The team leaves with a map of where to start and what to keep away from AI.
One process, taken apart
The room names a recurring piece of work that it believes AI cannot do. We break it into steps, find the explicit, format-conversion work AI can take on at each step, and mark where judgement stays with a person. In September the room chose the project budget process, the piece of work it was most certain AI could not touch. We broke it into six steps and found format-conversion work AI could take on at every one, while the estimating, the finance decisions and the approval stayed with people.
Agents in plain words, and one experiment each
What an agent is, the four things that get called one in Microsoft 365, and which to try first. We close by asking each person to name one experiment for the next two weeks, and we write down what the team agrees about sharing the results.
What your team leaves with
- A sorted map of the team's recurring work, exported from the board
- One of your own processes broken down step by step, with the AI opportunity and the human check marked at each step
- A named experiment for each person who wants one, and whatever the team agrees about sharing the results
- A written follow-up note within a week of the session
Price and practicalities
What it costs
£1,999 + VAT for the three-hour session, delivered remotely, for the whole team. The price includes a 30-minute preparation call with John before the day, so the examples use your tools and your reporting formats, and the written follow-up note afterwards.
The session is delivered by John McIntyre, HotPMO’s founder. He ran our AI for PMOs webinar in July 2026, from which this material comes, and presents the monthly PMO News segment on the PMO HotHouse livestream with House of PMO.
How it runs
- Three hours, delivered over Microsoft Teams into your meeting room or to a fully remote team
- The price is for the whole team, not per person
- Nobody needs to prepare anything. Bring a laptop and your approved AI tool. The session uses Mentimeter for polls and Miro for the shared board
- We have run it for a group of more than 20. For larger groups, ask us
- No client data, personal data or commercially sensitive material is used in any exercise
- If you want us in the room, ask and we will quote for it
Questions we are asked
Do we need Microsoft Copilot?
The session is built around the tools your organisation has already approved. Most teams we work with use Microsoft 365 Copilot, and the briefing structure and the agent examples are Microsoft’s. The two questions for placing a task, and the way we take a process apart, work with any assistant.
Is this training?
It is a working session. The team works on its own reports, logs and processes, and leaves with decisions and experiments written down. If you want accredited PMO courses, we would point you to House of PMO.
What about people who have never used AI at work?
They are usually a good part of the room, and the first hour is built for them. The board exercise asks for tasks, not technical knowledge, and the short demonstrations, such as sorting an inbox or scheduling a prompt, give beginners something to use the next morning.
How do you handle confidentiality?
Nothing personal, client-confidential or commercially sensitive goes into any AI tool during the session, and we ask for nothing of that kind in preparation. We work from generic scenarios and from formats you choose to share. HotPMO is ISO 27001 certified.
Can the session run in person?
Remote delivery is the standard and the price above assumes it. If you would rather have us in the room, tell us where and we will quote.
What happens after the session?
You receive a written note within a week: the board export, the process breakdown and the list of experiments. What the team does with them is up to the team. If you want help with the experiments, or a wider look at the PMO, the SPACE audit is the usual next step.
Where it leads
Three hours on AI tends to surface bigger questions about reporting, governance and how the PMO is seen by leadership. The SPACE audit answers those across six dimensions and produces a 12-month roadmap. The working session is the smaller first step.
Book a call about the session
Tell us about the team, the tools you have approved and what prompted the interest. We come back within one working day.
What happens next
- You fill in the form
- We arrange a call of about 30 minutes to check the session fits your team
- We agree a date and confirm the price in writing
- In the week before the session, a 30-minute preparation call with John to collect your reporting formats and examples
- The session, then the written note within a week
Explore more from HotPMO
The thinking behind the session is written up on the blog, with three free experiment guides, and the audit and SuccessHub pages describe where it can lead.