Harmonized Intelligence, PMI's concept for balancing people and AI in project management

When was the last time you stopped to think about whether the task you just delegated should have gone to a person or to an AI tool?

I'll admit I ask myself that question with a frequency that still bothers me. I automate generating status reports and feel like I've gained time. On the other hand, I know that in a negotiation with a stakeholder, an AI would hardly solve that for me. The thing is, between those two extremes, there's a gray zone, and that's exactly where most project managers' decisions happen. PMI gave that gray zone a name: harmonized intelligence.

The idea came out of the PMI Portugal 2026 Annual Conference, held on October 17 and 18 at Porto Business School. The event's theme was "Intelligence," under the motto "intelligence isn't just about knowing more, it's about leading better." It was within that theme that Lissa Muncer, from PMI's board, presented the concept.

What Harmonized Intelligence is

Harmonized Intelligence is the concept presented by PMI that describes the balance between human intelligence and artificial intelligence in project management, placing people at the center of the transformation and AI as support for decisions, not as their replacement.

Muncer's proposal isn't about which AI to use or how much to automate, it's about balance. Harmonizing, in this sense, means putting human intelligence and artificial intelligence to work together, with people at the center of the transformation, not as spectators of it. AI comes in as a way to strengthen capabilities and improve decisions, not as a replacement for human judgment.

I like this framing because it shifts the question from "how much do I need to automate?" to "what needs to stay human?" That second question is the one every project manager and PMO will need to answer more clearly in 2027.

How this connects to PMI's Talent Triangle

The conference's first day was organized around PMI's Talent Triangle, and the sessions covered a wide range: innovation, product management, AI, communication, branding, organizational culture, agile methodologies, PMO and emotional intelligence. It's no coincidence that emotional intelligence shows up next to AI in the same list of competencies.

On the second day, the "PM Group Quest" activity, led by Eduardo Espinheira, reinforced the same point from another angle: a collaborative challenge based on teamwork and joint problem-solving, showing that collective intelligence still decides a project's success and a tool hardly replaces that.

The message from the Talent Triangle, read alongside Harmonized Intelligence, is direct. Technical and process skills will keep mattering, but they're no longer the differentiator. The differentiator lies in the competencies AI doesn't replicate well: negotiation, reading context, emotional intelligence, judgment under ambiguity.

What changes in how you build a team in 2027

This has a practical consequence for anyone building or developing project teams. For a long time, we developed people around their ability to execute processes: build a schedule, consolidate a status report, run an agile ceremony. Much of that should already be supported by AI.

What needs to grow on the team is the ability to mediate conflict, to translate a technical problem for leadership, to notice when a stakeholder is saying "yes" but thinking "no." On a team like the one I manage in projects, that means rethinking what we put into each analyst's development criteria. My observations reflect personal perceptions drawn from my own experience and context, and don't represent a formal assessment of any specific company.

Building a project team in 2027 isn't about choosing between people and AI. It's about deciding where each one goes to get the best out of its capabilities.

An exercise to put this into practice

Set aside 20 minutes and answer, without filtering:

Which of the team's tasks today are pure process execution that could already be supported by AI without losing quality?

Which tasks depend on reading context, relationships, or judgment that no tool replaces today?

Where am I, as a manager, still spending a person's time on a machine's task, and where am I routing to AI something that would need a person judging the context?

Is our infrastructure, data, specifications and foundation ready to be used by both?

If I had to develop one competency on my team over the next six months to be better prepared for 2027, what would it be?

There's no right answer here. The value is in asking the question before efficiency pressure decides it for you.

If this topic interests you, I wrote before about project leadership and what technology still doesn't do for us, which pairs well with this reflection.

Best, Leandro | PMP | PMO and Project Manager