Practical guide
AI fluency for project managers: check a dependency plan
Build a fictional project schedule that distinguishes parallel work, prerequisites and shared-resource constraints.
Project managers can use AI to draft a schedule while checking the dependencies that make its dates possible. This original exercise plans a small internal guide launch using explicit task durations. It produces a dependency-aware schedule, not a delivery commitment or an estimate based on real team performance.
Define the tasks and their prerequisites
A date list should follow the work logic rather than substitute for it.
The fictional project has four tasks. Drafting takes two workdays. Design takes three workdays and can start only after drafting. Legal wording review takes one workday after drafting and can run alongside design. Final assembly takes one workday after both design and wording review are complete.
These durations are exercise assumptions, not professional estimates. Ask AI to represent the prerequisites explicitly. Do not allow assembly to start after design alone if wording review remains incomplete. Keep task duration distinct from a promised finish date and avoid inventing weekend or holiday rules.
Calculate the earliest finish under the stated assumptions
Parallel tasks contribute their longest dependent path, not the sum of every duration.
Use workdays numbered one onward, with tasks completing at the end of their final day. Drafting occupies days one and two. Design occupies days three through five. Wording review occupies day three. Assembly can take day six because both prerequisites are complete by then.
The assumed earliest duration is six workdays, not seven from adding all four durations. Show the two paths: drafting-design-assembly takes six days, while drafting-review-assembly takes four. This arithmetic does not establish that the team can actually perform the tasks in those durations.
Check whether parallelism requires the same person
A dependency diagram can omit the resource conflict that invalidates its dates.
Now change the scenario so the same specialist must perform both design and wording review, with no simultaneous work. Those four days of specialist work must be sequenced after drafting. Assembly then takes the seventh workday under the unchanged durations. Either specialist-task order has the same total here.
Ask AI to revise the schedule and explain why it changed. Do not preserve the six-day finish by assuming overtime or another specialist exists. If an additional resource is a possible recovery option, label it as requiring confirmation and approval rather than silently adding it to the plan.
Distinguish a delayed prerequisite from every downstream task
A change should propagate only where the dependency or resource relation requires it.
Return to separate resources, then extend wording review from one day to four. It now occupies days three through six and finishes after design. Assembly moves to day seven. The review path becomes the longer path; design does not need to start later merely because review takes longer.
Keep a short change note showing the altered assumption and affected finish. AI can recompute a plan, but inspect whether it shifted unrelated tasks or retained an obsolete dependency. More detailed formatting does not make an impossible sequence feasible. The explanation should remain understandable without a scheduling application.
Deliver the schedule with its conditions
The handoff should show what must remain true for the proposed dates to hold.
Include tasks, prerequisites, durations, resource assumptions and the three scenario finishes: six days initially, seven with the shared specialist and seven with the longer review. Mark them as calculated scenarios rather than commitments. Name the owner who can confirm duration and resource availability before real use.
Review the worker’s reasoning about dependencies and parallel work, not confidence in the first answer. This project exercise addresses delivery sequencing. The programme guide instead asks whether several workstreams jointly create a benefit, which cannot be settled by completing a schedule alone.
Sources and scope
NIST: AI risk management. Skills England: workplace AI foundations.
These references provide background, not validation or endorsement of this exercise. The case details, calculations and suggested review questions are original instructional material. Use them to discuss observable work, not to infer customer outcomes, professional credentials or performance in every setting. Before adapting the exercise, confirm the relevant facts, approved tools, data permissions and decision owners. If you change the case, revisit the expected answers and checks as well. These examples describe practice tasks, not a promise that a particular product includes the fictional features.
Sources: [1] [2]
Sources
- 1.AI RMF Core · NIST
- 2.AI foundation skills for work benchmark · Skills England