AI & Project Management
From Plato to Microsoft Project: Managing Projects in the Age of AI
How the Theory of Forms explains the gap between your project plan and reality — and what changes when AI agents join the team.
More than 2,000 years ago, Plato proposed the Theory of Forms: the idea that what we experience in the real world is an imperfect representation of an ideal model.
That is surprisingly close to how project management works.
A Microsoft Project plan is a model of how we expect reality to unfold. We define tasks, durations, dependencies, resources, calendars, costs, milestones, and baselines. The scheduling engine calculates the resulting plan.
But the Gantt chart is not reality.
It is our model of reality.
And artificial intelligence is making the relationship between the plan and reality much more interesting.
When AI Becomes a Project Resource
Traditional Microsoft Project schedules primarily assign work to human resources and equipment.
But imagine a project containing:
Humans and AI agents can now contribute work toward the same project objective.
But they operate very differently.
A human may work eight hours per day, have an hourly cost, take vacations, and execute a limited number of assignments simultaneously.
An AI agent may operate continuously, complete some assignments in minutes, execute multiple processes concurrently, and generate costs through tokens, API calls, models, and compute.
How do we estimate, schedule, cost, assign, and track work when project resources include both humans and AI?
A new resource model for Microsoft Project
Microsoft Project already provides a useful foundation through Work, Material, and Cost resources. AI could extend this model conceptually into four categories:
The project plan can therefore evolve from describing simply who performs the work to describing what combination of human intelligence, artificial intelligence, and computational capacity performs it — and at what cost.
AI work doesn't behave like human work
Consider a task: analyze 10,000 customer conversations and identify the top 20 customer intents. A human team might require weeks. An AI agent could potentially perform the initial analysis in hours. But faster execution doesn't necessarily mean the task is complete. The AI result may require human validation and another iteration:A typical AI task cycle — not a single hand-off, but a loop.
That changes project estimation. Instead of estimating only human work hours, project managers may increasingly estimate cycles of AI execution and human validation.
AI projects add another layer of uncertainty
Building AI systems introduces its own challenge. AI development often involves experimentation:Experimentation is the default, not the exception.
A model may fail to produce an acceptable result. An agent may require different instructions. Data quality may change the entire approach. A good AI project plan should therefore represent experimentation rather than pretending every activity will succeed on its first attempt.
Microsoft Project's scheduling principles — dependencies, resources, costs, milestones, and baselines — remain valuable. What changes is the nature of the work being modeled.
From project manager to AI conductor
From project manager to AI conductor As AI performs more project work, the role of the project manager also evolves. The PM increasingly coordinates human specialists, AI agents, automated workflows, models and data, and computational resources. The project manager becomes an AI Conductor. Four skills become especially important:
Microsoft Project can provide the underlying structure through which that work is organized.
Bringing AI to Microsoft Project and MPP files
AI can also change how project managers interact with existing project schedules. Instead of manually navigating a complex plan, imagine asking:Ask about your existing plan
| You ask Erix |
|---|
| Why is my project late? |
| Which tasks threaten the deadline? |
| What changed from the baseline? |
| Which resources are overloaded? |
| Create a recovery plan. |
| Update my project from this week's status information. |
| Generate an executive status report from my MPP file. |
Ask about delegating to AI agents
| You ask Erix |
|---|
| Which tasks could be delegated to AI agents? |
| What will their execution cost? |
| Which AI assignments require human review? |
| How should I rebalance work between humans and AI? |
This is the direction we are exploring at Housatonic Software with Project Plan 365 and Erix: combining AI with Microsoft Project workflows and MPP files.
The intelligent project plan
Plato distinguished between the ideal model and its imperfect manifestation in reality. Project management faces the same challenge.The plan represents what should happen. Project execution reveals what actually happens.AI can help continuously connect the two — analyzing the project, identifying deviations, explaining risks, recommending corrections, generating reports, and helping update the plan. The result is something more powerful than a static Gantt chart: a living model of work performed by humans and AI together. The future of AI for Microsoft Project may therefore be about much more than adding a chatbot to project-management software. It may be about creating the architecture through which organizations estimate, schedule, cost, assign, monitor, and coordinate human and AI work together.
Building the Instruments
At Housatonic, we are building the instruments and the platform that help Project Managers become better AI Conductors—through Project Plan 365 and Erix AI. Our mission is simple:
We are building AI to amplify the Project Manager.
