When I facilitated the University of New Hampshire Professional Development & Training class “AI in Project Management,” participants’ questions offered a useful window into how project managers are approaching the technology.

They asked everything from “What can you use AI for?” to how AI could help compare current projects with previous ones. Other questions focused on prompting, defining scope, scheduling deliverables, implementing workflows, and tracking tasks.
In other words, could AI help manage the project, preferably before the next status meeting?
What interested me was what these questions suggested about participants’ experience and how project managers might approach using AI.
Familiar Work, Unfamiliar Tools
The questions connected AI to recognizable project-management responsibilities. Participants were exploring how the technology might fit into work they needed to accomplish.
That does not mean everyone had the same level of experience. Asking how to use AI to build a schedule does not necessarily indicate limited scheduling knowledge. The unfamiliar part may be communicating requirements to an AI tool or evaluating its response.
Project-management experience and AI experience are different. Someone can confidently lead a project while still learning to prompt. Another person might use AI comfortably but need help recognizing weaknesses in a project plan. Being able to generate a beautiful table does not mean the people listed in it are available on Tuesday.
The questions alone cannot establish anyone’s proficiency. They do suggest that knowing how to use a tool and knowing how to manage the work deserve separate attention.
From Generating Output to Applying Judgment
“What can you use AI for?” explores possibilities. “How can I compare my project with previous projects?” identifies a specific application.
Questions about implementing workflows raise another issue: what happens after AI produces an answer?
Suggesting action items is different from assigning them to people. Drafting a schedule is different from committing a team to deadlines. Your AI may be enthusiastic about a Friday launch. Your team may have questions.
The approach I favor puts judgment at the center. Ala the “Halo Effect”, a polished response is a draft to examine, not evidence that the underlying plan is sound. Project managers still need to decide whether the output reflects the project’s actual circumstances and is ready to use.
General Guidelines for Using AI
-
Start with a real project need. Identify what you want to improve before choosing an AI feature. “We should do something with AI” is enthusiasm, not a project objective.
-
Choose work you can evaluate. Begin with a familiar task whose errors you are equipped to recognize.
-
Provide context. State the relevant constraints and identify the information the AI should use. “Make me a project plan” leaves rather a lot to the imagination.
-
Separate assumptions from commitments. Treat suggested dates and assignments as proposals until reviewed.
-
Define responsibility for review. Decide who will verify the output before it enters the project’s working records.
-
Establish boundaries. Clarify what information may be shared and which actions require approval.
-
Measure the full effort. Include preparation and correction time. A ten-second draft followed by two hours of repairs deserves an honest review.
-
Learn from each experiment. Record what worked and adjust your approach before repeating it.
A Final Thought
Approach AI as assistance, not a transfer of accountability. Give it useful work and examine the results before acting. And before accepting that impressive schedule, check whether it has quietly assigned someone a 47-hour Tuesday.


