Artificial intelligence is reshaping the project management landscape, but for many PMOs, making sense of its real-world impact can be tricky. Almost every discussion we have right now at events, with customers or speaking to partners is the topic of AI in PMO comes up. But what we keep seeing is that everyone has questions, but they’re still looking for actual answers!
To bring clarity to this rapidly evolving space, we’ve gathered insights from leading experts across Altus partners in a special roundtable discussion. In this article, you’ll discover practical perspectives and actionable advice designed to help PMO leaders cut through the hype and confidently navigate the opportunities and challenges AI brings to their everyday work.
How can organisations assess and understand their readiness to embrace AI?
The first step is to take a hard look at your current state. Do you have clean, accessible data? Are your processes consistent enough to benefit from AI?
Beyond technology, readiness is about people. Helping them do things differently and even think differently about how AI can support everyday work. Does your team have the skills and mindset to embrace AI, and is leadership ready to support it?
A quick assessment across data, processes, technology, and people will reveal where you are and where to focus first. AI works best when the organization is prepared, not just when the tool is powerful.
Kelvin Kirby (Technology Associates):
Organisations need to stop guessing and take a sober look at four things: Data hygiene and availability; process maturity; their technology landscape; and the culture and adoption appetite.
If data hygiene is lacking, the portfolio/project data is a mess—inconsistent schedules, incomplete resource data, weak financials—AI will only amplify the chaos. A readiness check starts with data quality scoring across projects, resources, and financials.
When you look at maturity, if every team manages projects differently, you don’t have a foundation AI can reliably automate or optimise. Assess against a framework (P3M3, CMMI, or your internal methodology). Low maturity = limited AI benefit.
You also need to consider your technology, evaluate whether your PPM platform (Project Online, Altus, Project for the Web, Planview, Clarity, etc.) can expose APIs, integrate with automation tools, and handle AI-generated insights. Legacy customisations or siloed tools slow everything down.
Finally, you need to look at your culture and the appetite for adoption. If project managers are defensive about scrutiny or PMO is seen as a policing function, AI adoption becomes political. Assess trust, openness to change, and leadership alignment.
Ryan Darby (Sensei Project Solutions):
You can assess and understand your readiness for AI by not thinking of it as AI. Don’t think of it as some new miracle technology. Focus on existing problems and inefficient processes. Work out how to improve them and then see how AI can contribute to that solution. Everyone wants AI, but can’t articulate why or what it should do. You don’t need to worry about readiness until you go through a process of understanding your requirements and then determining which AI is appropriate where, and what the value will be. If you can’t identify processes that can be improved with any technology then you can’t determine where AI can play a part.
In addition, look at your data. Do you have any in a clean state? And is your security in your Microsoft and other environments setup properly?
A good way for organizations to gauge their AI readiness is to take a close look at the foundations they already have in place: their data, their processes, and their culture. AI works best when data is reliable and accessible, and when processes are clearly defined. Just as important is whether teams are open to trying new approaches and learning as they go. A quick assessment of data quality, technology infrastructure, governance, and workforce skills can give organizations a clear picture of where they’re strong and where they need to improve before investing heavily in AI.
It also helps to understand the “why” behind AI adoption. Organizations that connect AI initiatives to real business goals—whether improving decision-making, increasing productivity, or reducing manual effort—tend to see better outcomes. Clarity of purpose makes it easier to prioritize the right use cases and build momentum.
What’s one action PMOs should take today to start their experimentation and journey with AI?
Tim:
Start small but start smart. Pick one area that can deliver real value without causing disruption if it fails. This could be risk forecasting, resource planning, or automating repetitive reporting.
Run a pilot, learn from it, and share the wins. These small experiments build confidence, teach the team, and create momentum for larger AI adoption. The important thing is to get moving. AI is not a future problem; it is something PMOs and organizations can start exploring today.
Kelvin:
Pick a single low-risk, high-annoyance problem and prototype an AI-assisted workflow around it. This could be:
- Automatically summarising weekly status reports;
- Drafting RAID entries from meeting notes;
- Identifying late tasks and potential slippage; or
- Flagging resource overallocations and suggesting options.
Most importantly, don’t form a committee, and don’t write a strategy deck that gathers dust. Just pick a use case, build a prototype, measure the impact, and iterate.
Max:
A practical first step for any PMO is to pick one everyday process that creates friction and test a small AI use case around it. This could be something like automating status updates, summarizing risks and issues, or using generative AI to clean up project documentation. Starting with something simple and contained helps teams learn quickly, demonstrate value, and build confidence without taking big risks.
This first experiment also gives the PMO insights into how AI fits into existing workflows—what data is needed, how results should be validated, and how project teams interact with the technology. Those early lessons become the building blocks for a broader, more sustainable AI strategy.
Ryan:
Today? Prepare your data. Start now, even before you understand what AI is.
Most PMO’s do not have their data in a clean and neat and consistent format now. This is something they need to fix for good reporting and is also an essential first step for AI. AI needs data be it past projects or reference documents such as policies and procedures to reference.
While you are working out what to do, get on with cleaning your data. You need a sample set of projects that AI can use.
What are the most exciting use cases (current and potential) you’ve seen for AI in the PMO and PPM space?
Tim:
AI is already making a difference in predictive analytics for project risk, smart resource balancing, and portfolio prioritization. But the exciting part is what’s coming next. Imagine AI helping plan projects dynamically, modeling multiple scenarios in real time, or even preparing governance reports automatically.
Too often, organizations work hard to compile data and update tools, yet have minimal capacity to analyze it. The “Minority Report” movie helps illustrate the kind of real-time, interactive analysis that should be possible for leaders and end users, asking questions and immediately using results to explore new questions in a creative flow.
Today, there is too much friction in data analysis, and next-level AI-driven insights can empower PMOs to make faster, smarter decisions. These tools do not replace humans. They free PMOs to focus on strategy, decision-making, and delivering real business value instead of getting bogged down in spreadsheets.
Kelvin:
Here are the ones that actually move the needle. Right now, we can already:
- Run automated health checks that pull data from PWA or your PPM tool;
- Schedule quality scoring and recommendations (dependencies, constraints, baselines);
- AI-driven resource utilisation analysis (overallocations, underuse, forecast variances);
- Status report generation from timesheets, updates, and risks; and
- Translate meetings into actions (Teams or Zoom → RAID updates and tasks)
In terms of where things are headed, I think some of the exciting and emerging use-cases are:
- Predictive forecasting for cost, schedule, and resource slippage
- Adaptive portfolio prioritisation based on business outcomes, not spreadsheets
- AI QA for plans — catching inconsistent work breakdown structures before approval
- Cross-project dependency mapping driven by NLP, not manual detective work
- Personalised PM coaching assistants embedded in PPM tools
Max:
Some of the most exciting AI use cases in the PMO and PPM space are the ones that remove friction from everyday work and allow teams to focus on higher-value activities. For example, automatic project setup can take what is normally a time-consuming administrative task and streamline it into a guided, AI-driven flow—automatically creating templates, schedules, risk logs, and governance checkpoints based on project type. AI can also turn meeting transcripts directly into entries in the PPM tool, automatically updating actions, risks, decisions, and next steps without the project manager needing to rewrite everything afterward.
AI is also transforming follow-up and compliance. Tools can now nudge team members for timesheet completion or status report updates, reducing the amount of manual chasing PMOs typically do. Even more powerful is the ability to identify governance violations in real time—for example, detecting that a project is trending toward a cost overrun but no change request has been raised. These kinds of “always-on” controls help PMOs maintain quality without increasing workload.
At the same time, predictive analytics and portfolio optimization continue to be major areas of excitement. AI can analyze risks, capacity, and financials to spot issues earlier and recommend a better mix of initiatives. Combined with generative AI to automate reporting and documentation, these capabilities free PMOs and project managers to focus more on strategy, stakeholder alignment, and delivering value.
Ryan:
Don’t worry about it. Look at where you are experiencing pain first. Clients are making a mistake by rushing to shiny solutions vendors show them. Having said that, Sensei is working with a group of clients to understand pain points and prioritise use cases. The most common so far relate to the reduction in administrative effort, especially capturing data from meeting minutes and then turning that into data and updates in Altus. This is an interesting use case as when we have spoken to so called PPM system expert they all jump to what they think a use case is, and don’t consider the actual time consuming work of a project manager which is largely about transposing data. In addition stand out use cases relate to lessons and risks and how they can be identified on new projects. This is also the most popular demo we do for AI.
In short don’t look for new problems, look at the old annoying problems and use AI to help with them.
In addition longer term think of what entirely new features that AI could add to a PPM solution. The strategy area is a key one for this, as AI could be the driving force behind this as nothing much exists in this space anyway.
What’s the biggest question about AI for PMO and PPM that is yet to be answered?
Tim:
The biggest challenge is balancing AI with human judgment. AI can provide insights, but PMOs still need to make decisions that reflect experience, context, and organizational priorities.
How do you build trust in AI recommendations and integrate them into governance without losing accountability? That question will define sustainable AI adoption in PMOs. Part of the PMO lifecycle involves standardizing tools and processes, learning from the data, and continuously tuning and improving both methodology and technology.
A strong tool helps, but its value is maximized when leaders are educated, engaged, and committed to improving processes. That combination of smart leadership, strong methodology, and AI-driven insight allows PMOs to ask the questions they could not answer before and solve them faster.
Kelvin:
For me, everyone loves the idea of dashboards and “smart suggestions,” but the real test is this:
- Will the PMO allow AI to auto-correct a broken schedule?
- Will resource managers let AI adjust allocations?
- Will portfolio boards accept prioritisation recommendations without politics overruling them?
The unanswered question isn’t a technical one. It’s whether businesses will give AI actual authority rather than treating it like another reporting tool.
Ryan:
What is the killer use case?
So far everyone is looking for one but no one is trying to solve actual problems. And that will lead to the question of, what will a PPM solution look like in the future? It won’t look like Altus, not in the shape we see it now.
Max:
The biggest open question is how PMOs should govern AI as it becomes more embedded in decision-making. We still don’t have a clear, universal framework for ensuring AI-generated insights are transparent, trustworthy, and ethically used. Questions like: How much should we rely on AI predictions? Who’s accountable when AI influences key decisions? What standards should we set for quality and oversight? are still being worked out.
Until those answers mature, PMOs will need to balance innovation with responsibility. They’ll play a key role in setting guardrails that encourage experimentation while protecting data integrity, maintaining transparency, and ensuring humans stay firmly in the loop.
