We have been running practical AI training for Project Management Offices (PMOs) and project managers for around 18 months. Over that time, those of us delivering the training have developed a strong view of how PMOs are actually using these tools, along with where the real opportunities for improvement sit.
A consistent pattern emerges after training. Teams return to their organisations and quickly realise what is possible, often well beyond what they had previously considered.
This article explores key PMO functional areas where these tools can be applied, and how the PMO operating model can be reshaped to scale the benefits.
Governance and Reporting
Governance and reporting has long been a source of frustration for PMOs and project managers.
Most teams produce multiple reports each month for different audiences such as sponsors, steering committees, boards and external stakeholders. While the underlying information is largely the same, each version requires a different lens. Historically, this has meant rewriting similar content repeatedly.
This changes when a base report can be reused and reshaped for different audiences. With clear context and examples of how reports have previously been tailored, outputs can be reframed quickly and consistently.
In practice, this means:
- Less time spent rewriting reports
- Faster turnaround across stakeholder groups
- More consistent messaging
Human review is still required. In our experiments, large language models will return a successful or ‘true’ output 70-80% of the time. Outputs will typically land well but not perfectly. The problem is that AI tools will never say ‘I don’t know’, it will waffle or hallucinate and someone with the expertise to tell the difference between the two remains essential.
A simple example is reframing a report for external stakeholders by:
- Removing detailed financial information
- Simplifying technical risks
- Focusing on impacts, progress and upcoming changes
When prompts and examples are reused, this becomes increasingly efficient. Even this single use case delivers a significant productivity gain for PMOs.
Meeting Minutes and Transcription
Meeting transcription and minutes is one of the most widely adopted use cases and one where tangible benefits are realised quickly.
Meeting outcomes can be recorded, transcribed and converted into structured minutes with minimal effort.
AI tools will:
- Record meetings
- Transcribe discussions
- Generate structured minutes
- Distribute them automatically
Some organisations have already moved to a fully automated, no-touch model, where the system will transcribe the meeting, do the minutes and distribute the minutes at a preset time. While this can work in low-risk environments, we would always recommend a ‘light-touch’ approach rather than a ‘no-touch’ approach.
The value of human review is not just accuracy. It is about emphasis. AI can miss nuance, context or the importance of specific decisions and actions. A quick review allows you to:
- Reinforce key messages
- Adjust tone or emphasis
- Add clarity where needed
- Pick up any errors or inconsistencies
The productivity gains here are substantial. Tasks that previously took one to two hours per meeting can now be reduced to a few minutes of review.
It also removes a low-value activity that very few people enjoy.
Risks and Issues Management
Risk and issue management is another area where outcomes can materially improve.
Rather than focusing on whether a register exists and has been updated, attention can shift to quality. Risk descriptions, mitigations, assumptions and review cycles can be challenged and improved.
For example, a register can be reviewed against questions such as:
- Are risks described clearly and completely?
- Do mitigations align to causes rather than symptoms?
- Are likelihood and consequence ratings credible?
- Has the register been meaningfully updated?
This supports better project control. The effort is refocused from administration to analysis, with registers becoming more useful as decision support tools.
Schedule Analysis
Professional scheduling tools remain essential for building and managing schedules. These tools are designed for structured data, dependencies and forecasting.
The opportunity here is not replacement, but assurance.
Schedules can be reviewed against recognised best practice, quality indicators and delivery risks. This includes identifying:
- Weak logic or dependencies
- Unrealistic sequencing
- Areas of schedule risk
- Gaps against agreed standards
For complex projects and programs, this can significantly reduce review effort. However, data quality matters and large schedules often need to be reviewed in sections. Expert oversight remains essential to interpret outputs and identify errors.
Stage Gate Reviews
Stage gate reviews are often resource intensive without always delivering proportional value.
Significant effort is spent chasing artefacts, reviewing submissions and assessing whether documentation meets required standards. Feedback cycles can be slow and frustrating for both PMOs and delivery teams.
This process can be streamlined using large language models. Artefacts can be reviewed quickly against defined criteria, with gaps and weaknesses highlighted early. This allows the PMO to focus on assessment and analysis rather than basic compliance.
Instead of waiting for feedback, project managers can effectively run a ‘pre-stage gate’ review themselves, through a dedicated agent or pre-defined prompt. This proactive approach can positively impact team behaviour. Rather than submitting an artefact and hoping it passes, teams can iterate upfront. The quality of submissions improves and the back-and-forth reduces, saving time and keeping stress levels down.
The result is fewer rework cycles, less friction and a more efficient assurance process overall. It also supports a shift in how the PMO is perceived, from policing artefacts to maintaining meaningful standards.
Lessons Learned
Most organisations are not short on lessons learned. It’s actually using the lessons where organisations tend to fall short in practice. While lessons are commonly captured at project close, they are rarely visible, searchable or embedded into future delivery. They exist, but they are not actively used.
Lessons can be strengthened by prompting teams to be more specific about context, impact and future application. More importantly, they can be reintroduced earlier, before launching new initiatives, entering complex phases or designing governance and delivery approaches.
This shifts lessons learned from a retrospective exercise to a practical input into better decisions.
Portfolio Management
At a portfolio level, analysis and decision support often rely on static outputs generated through PPM systems and spreadsheets. These tools may remain important, but interpretation and rework of the outputs can be time consuming. There may also be differing and subjective views on ranking of initiatives and even how the ranking should be done. With some initial effort put into setting parameters, AI tools can help support standard inputs and outputs.
Portfolio data can be reviewed more quickly to identify patterns, exceptions, interdependencies and risks. The same information can be framed differently for executives, steering committees or delivery teams. This reduces manual effort, speeds up reporting and supports clearer discussions around priorities, trade-offs and capacity.
At a portfolio level, AI tools can improve both the speed and quality of analysis, reporting and decision support.
Reshaping the PMO Operating Model
Across these use cases, a consistent pattern emerges. Effort shifts away from repetitive administration and towards analysis, insight, assurance and decision support.
To realise these benefits at scale, PMOs need to move beyond ad hoc use. This means defining standard use cases, developing reusable prompts and workflows, embracing agentic tools and setting clear expectations for review and embedding these practices into everyday PMO activities.
The objective is not to replace professional expertise. It is to reduce time spent on low-value tasks, improve consistency and free people to focus on the areas where they add the most value.
For PMOs that do this well, the impact goes beyond incremental efficiency. It represents a fundamental shift in the role and contribution of the PMO.
Can we help?
If you’re looking to establish a PMO operating model with AI at its core, or would like support training your team on practical project management and PMO use cases, get in touch today.
