Consultant With Decades of Experience Shares Why AI Should Be a Strategic Thinking Partner, Not a Project Driver
Mike Hannan is the founder and principal consultant at Fortezza Consulting. With decades of experience coaching project leaders at every level, he has spent his career helping organizations navigate complex delivery systems, optimize portfolios, and challenge conventional wisdom.
In this conversation, Mike shares how he’s using AI, not to replace project practices, but to push teams to ask better questions, avoid false certainty, and get closer to the real levers of performance.
Understanding System-Level Performance Drivers Matters More in the AI Era
I'm the founder and principal consultant at Fortezza Consulting, where I coach leaders at all levels to improve project portfolio performance. That means helping them think more systemically, especially in environments shaped by constraints, interdependencies — and now, AI.
Unlock for Free
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Already a member? Log In
In my opinion, in an AI-first world, the need is stronger than ever to truly understand what’s driving performance at the system level. That's still the only way to navigate complexity better than the competition.
AI can't do that yet. And it won’t get there until it’s better trained to separate garbage from gems.
How to Use AI for Project Templates, Delivery, and Portfolio Optimization
That said, project delivery is changing. I find the need for most project templates has gone way down. Most large language models can now generate solid templates and even help tailor them for specific circumstances.
For example, Claude does a pretty good job of this. I might prompt it with:
- “Please give me a template for a project charter tailored for the cosmetics industry.”
- “Please give me a template for a risk register tailored for the nuclear-energy industry.”
From there, I’ll refine it further: “Focus on IT projects in cosmetics,” or “Focus on nuclear waste disposal risks.” The responses are quick, flexible, and surprisingly accurate.
I’m also really encouraged by emerging AI capabilities that can evaluate hundreds or even thousands of scenarios to find the optimal solution under real constraints — for example, identifying the highest-value project baseline.
One case I like comes from Lightning Motorcycle, which used Autodesk’s AI-enhanced CAD tools to design far more optimal components than human engineers could design alone.
Another is TransparentChoice, which uses AI to help clients find the staggered-flow model that maximizes throughput within resource constraints. It simplifies a tough question: “How many projects can we take on and still achieve the highest possible completion rate?”
Where AI Adds Value in Project Delivery — and Where Humans Are Still Essential
Rote tasks like meeting notes, key takeaways, and synopses of large bodies of content are also handled very effectively by AI. That alone gives me back several hours each week.
For those, I mostly use Otter.ai, which is my preference for capturing and summarizing discussions. But tools like Copilot work well too.
The areas that still demand a human touch are varied. Anything in the social or human services space—where mentorship, empathy, or trusted relationships matter—still depends heavily on people. The same goes for anything requiring deep system-level thinking or critical judgment.
Those are still human domains.
How to Integrate AI Into Complex Workflows While Keeping Human Judgment Central
Sometimes I take AI further and set up a model that defines everything needed to synchronize work across people, functions, and even supply-chain partners.
Because AI can be slotted into that model wherever it makes sense. Humans then need to adjust based on how the mix of AI and human agents is performing.
Here's a simple analogy: Imagine an executive assistant retires, and the executive is distraught about losing that finely-tuned support. Instead of hiring a replacement, you build a model that encodes preferences:
- “Always 15 minutes between meetings”
- “Prefer airline A unless no nonstop is available”
- “Focus time can only be interrupted if exec A or customer B insists”
Over time, the model refines and covers 95% of scenarios well.
But there’s always a point where human judgment is indispensable — reading emotional cues, cultural nuance, or shifting priorities. What if a family visit isn’t welcome? What if a customer starts pushing too hard? What if a boss’s urgency is rising in subtle ways? That’s where the blend of human and AI becomes valuable.
In projects, the model is more complex, but the principle is the same: Build it, refine it, and orchestrate the mix of AI and human agents so each does what they’re uniquely effective at.
AI Doesn’t Truly Understand Complexity
While AI helps me spend less time summarizing or distilling information, it also creates a dangerous illusion: that the models themselves “understand” deep complexity. They don’t.
Because of that illusion, many clients are looking for faster results, hoping AI will deliver what they see as "quick fixes". To help them get there, I focus on flow-acceleration techniques grounded in the Theory of Constraints and Lean. Those approaches help projects move faster without cutting corners.
And when we talk about setting expectations, I usually encourage clients to aim higher — to expect more acceleration. Setting modest goals tends to produce modest results.
So my role today is less about summarizing or delivering takeaways, and more about helping leaders pause, step back, and challenge their assumptions. That’s where the real performance gains come from.
How to Use AI to Challenge Assumptions and Improve Critical Thinking
I use AI as a target; not a solution. It’s there to be challenged. It's a thinking partner, rather than a driver.
I use AI as a target; not a solution. It’s there to be challenged. It’s a thinking partner, rather than a driver.
A good example is taking the generic, surface-level answers that LLMs generate and using them as a target for critique. I’ll ask clients to pick those responses apart:
- “What would have to be true for this to be false?”
- “In what context might this be completely wrong?”
Those questions train them to engineer smarter follow-up prompts, which often lead to real insight.
And the questions sharpen their critical thinking too, which is particularly important now. It's surprising how susceptible most of us are to confirmation bias. When AI gives people an answer that supports what they already believe, they tend to feel validated — and then turn off their critical thinking.
That’s not just surprising; it’s scary.
Modernizing Project Delivery While Maximizing ROI
For me, any technique or approach that helps maximize the ROI of a project portfolio—instead of just managing baselines without looking for higher-value alternatives—is pure gold in my book. These aren’t always “lightweight,” but they do take some of the burden off people while increasing overall impact.
A few examples I use with clients—all of which will appear in the upcoming PMBOK 8th edition (which I helped author)—include Value Breakdown Structures, Critical Path Drag, and Critical Path Drag Cost.
These extend the traditional critical path concept to the value side of the equation. For example, if I can accelerate the critical path by a month but it costs $200K to do it, will the additional value outweigh the cost?
Other powerful approaches include:
- Critical Chain Project Management (CCPM) at both the project and portfolio levels — to maximize throughput and due-date performance.
- Analytic Hierarchy Process (AHP) — to select higher-value projects.
Each of these techniques is designed to maximize ROI, not just manage delivery.
Agentic Workflows Aren’t Worth the Effort — Yet
If AI could be trained to proactively surface higher-value alternatives as they arise—without being prompted—that would be a real breakthrough. That’s when AI would stop being just a tool and start acting like a true contributing member of the team.
I’ve started experimenting with this conceptually. There are some promising models emerging that could make agentic workflows more user-friendly and mainstream, but they still require a lot of upfront effort to train for the unique context of each organization.
Even something simple—like teaching an agent to understand my travel preferences and suggest full itineraries—has taken more time to train than it’s saved me so far. But once it really learns, I’m confident the payoff will flip and that same principle will apply to project delivery.
Why People Who Master AI Will Reshape Project Delivery
Ultimately, I don’t think AI will reshape project delivery. I think the people who learn how to master and orchestrate AI will reshape it.
Ultimately, I don’t think AI will reshape project delivery. I think the people who learn how to master and orchestrate AI will reshape it.
And very complex—even “wicked”—problems will start to produce simpler and more elegant solutions, generated faster than ever by the brute-force computing power of agentic AI.
Why Breadth-First Thinking Will Define the Next Generation of Delivery Leaders
My advice is simple: Breadth first.
Most of us were taught that specialists and technically deep experts will be rewarded financially more than generalists, and this has been true for many decades now. But the more that AI can address technically deep problems faster and more effectively than humans can, the more valuable end-to-end, breadth-first, system-level understanding will become.
The more that AI can address technically deep problems faster and more effectively than humans can, the more valuable end-to-end, breadth-first, system-level understanding will become.
After all, if you don't know enough about the full scope of the problem to orchestrate a network of AI agents to help you solve it, then your value in the job market will be diluted.
Follow along
You can follow Mike's work on LinkedIn and learn more about his consulting at Fortezza Consulting.
More expert interviews to come on The Digital Project Manager.
