The first time I watched an AI tool estimate project timelines, I felt a mix of excitement and skepticism. As a software engineer, I’ve seen enough Gantt charts and “optimistic” roadmaps to know that predicting delivery is hard—even for humans who know the codebase.
But as AI quietly moved into our project management stack, something interesting happened: instead of replacing project managers, it changed what great project management looks like.
I’m Phong Lee, and here’s how I see project management evolving in the age of AI, from the perspective of someone who lives on the execution side of those plans.
From guesswork to data-informed planning
Traditional project planning often relies on:
- Gut feel and past experience
- Manually updated spreadsheets
- Best-case assumptions under pressure
In one organization, we started feeding AI tools with:
- Historical ticket data (estimates vs. actuals)
- Commit and deployment history
- Incident and rollback records
The system began to surface patterns:
- Certain types of tasks consistently took 2–3x longer than estimated.
- Specific integration points were frequent sources of delay.
- Teams with fewer context switches delivered more predictably.
Instead of replacing project managers, AI became a reality-checking partner, helping set expectations based on evidence, not hope.
How AI actually helps project managers (from my seat)
Here’s where I’ve seen AI add real value:
- Risk prediction: flagging epics likely to slip based on scope, dependencies, and team history.
- Dependency mapping: highlighting work that’s blocked or blocking others.
- Capacity insights: showing when teams are overloaded before it becomes a crisis.
- Status summaries: auto-generating rollups from tickets, commits, and comments.
On one large initiative, an AI assistant flagged that our “simple integration” epic had a higher-than-normal risk score based on similar past work. That nudge led us to invest more time in discovery—and we uncovered hidden API constraints that would have derailed us later.
What doesn’t change: human judgment and communication
Even with AI in the loop, the most important project management skills I’ve seen remain deeply human:
- Asking good questions: “What are we not seeing yet?”
- Facilitating alignment across engineering, product, and business.
- Navigating trade-offs under real-world constraints.
- Protecting focus so teams can actually deliver.
On one team, our project manager used AI-generated risk reports not as the final word, but as a conversation starter:
- “The model thinks this epic is high risk. Do you agree? Why or why not?”
- “Historically, this pattern of changes leads to incidents. How can we de-risk?”
That blend of AI insight and human judgment was far more powerful than either alone.
New skills for project managers in the age of AI
From my viewpoint as an engineer, the best modern project managers:
-
Understand the data behind the dashboards
- They know what inputs drive AI predictions.
- They question outputs instead of accepting them blindly.
-
Design healthy workflows for humans and machines
- They ensure ticket hygiene so AI has quality data.
- They create processes where AI suggestions are easy to act on.
-
Communicate probabilistically, not absolutely
- They talk in ranges and confidence levels.
- They make room for uncertainty instead of hiding it.
-
Champion sustainable pace
- They use data to argue against impossible timelines.
- They show how overloading teams today hurts delivery tomorrow.
My experience inside AI-augmented projects
On AI-assisted projects, my day as an engineer changed in subtle but powerful ways:
- Standups were shorter and more focused because everyone could see live status and risks.
- Our project manager spent less time chasing manual updates and more time unblocking dependencies.
- We caught potential delays weeks earlier, when small adjustments still mattered.
I, Phong Lee, stopped seeing AI as a threat to project management and started seeing it as a tool that:
- Reduces manual status work
- Surfaces patterns humans might miss
- Gives us a more honest picture of where we are
Guardrails for using AI in project management
To make AI genuinely helpful (and not just another dashboard), enterprises should:
- Protect privacy: be thoughtful about what data is fed into models.
- Avoid surveillance metrics: don’t weaponize activity tracking against individuals.
- Focus on team-level insights: use data to improve systems, not punish people.
- Keep humans in control: project managers and teams make the final calls.
Used well, AI becomes a microscope and telescope for projects—helping leaders and teams see both the small details and the big picture more clearly.
The future of project management in the age of AI
In the coming years, I expect:
- Project plans that continuously adjust based on live data.
- More emphasis on systems thinking, less on manual task chasing.
- Project managers evolving into orchestrators of humans, tools, and data.
If you’re leading projects today, the goal isn’t to compete with AI. It’s to:
- Leverage AI to clear away low-value work.
- Double down on the human skills of communication, judgment, and leadership.
- Build environments where engineers and project managers can do their best thinking.
That’s what effective project management in the age of AI looks like from my side of the commit history.