Amplify Me: AI Toolkit for Project Managers

Replace the context with a short approved brief and name the person who accepts each deliverable.

Act as a project planning partner. Help me define this project before suggesting a delivery plan.

Context: [business problem, intended users, desired outcome, sponsor, constraints and existing evidence].

First identify missing information and ask up to five questions that could materially change the scope. If I cannot answer, label the gaps; do not invent facts, commitments, budgets or dates.

Using only the supplied facts, draft: problem statement; measurable outcomes; in-scope deliverables; explicit exclusions; constraints; assumptions requiring validation; and acceptance criteria for each deliverable. Distinguish outputs from the benefits they are meant to create.

Identify ambiguous terms, conflicting expectations and decisions the sponsor must make. Show how a proposed scope change would be assessed and approved. Finish with the smallest useful next step and an approval checklist. Treat this as a draft for human review, not an authorized project baseline.

Supply the actual working calendar and known availability; validate sequencing with delivery leads.

Act as a delivery planner. Break the agreed scope into manageable work packages and reviewable milestones.

Inputs: [approved outcomes, deliverables, exclusions, required dates, available team and working calendar].

Ask about any missing acceptance criteria, specialist skills or mandatory approvals before estimating. Do not invent task owners, capacity or fixed durations. Where estimates are necessary, use clearly labeled ranges and explain their basis.

Produce a table with work package, observable output, proposed accountable role, predecessor, acceptance test, estimated effort range and confidence. Separate effort from elapsed time. Include review, integration, testing, handover and operational readiness work.

Identify milestone exit criteria and the dependencies that could determine completion. Flag tasks that can run in parallel only if capacity permits. Show which dates are constraints versus provisional forecasts. End with sequencing decisions for the team and the evidence needed to turn this into an approved schedule.

Add your organization’s risk definitions and escalation thresholds before prioritizing.

Act as a constructive project risk reviewer. Review this proposed project and identify material risks, not a generic checklist.

Context: [objectives, delivery approach, dependencies, stakeholders, constraints and known issues].

Ask clarifying questions where missing facts would change the assessment. Distinguish a future uncertainty from an issue that has already occurred. Do not assign numerical probabilities without evidence.

For each risk, state cause, uncertain event and potential effect on an objective. Include a proposed owner role, early warning signal, prevention or mitigation, contingency and decision trigger. Cover delivery, people, integration, suppliers, security, adoption and operational handover where relevant.

Explain the basis of any qualitative likelihood and impact judgment. Mark assumptions and unknowns. Identify correlated risks and mitigations that could create new risks. Prioritize the few risks that need attention now and name the evidence or expert review needed. Finish with questions for the project team to validate the register.

Provide actual supplier commitments and interface acceptance criteria; keep proposed dates clearly labeled.

Act as a systems-minded delivery reviewer. Examine this work plan for dependencies and bottlenecks.

Inputs: [work packages, milestones, interfaces, external teams, supplier commitments, approvals and known capacity].

Identify missing dependency information before making claims about the critical path. Do not infer confirmed commitments from proposed dates. Separate internal sequencing, external dependencies and resource constraints.

Create a dependency register with provider, consumer, required output, acceptance condition, needed-by date if known, commitment status, proposed owner role and fallback. Flag cycles, single points of failure, long approval lead times and integration work that is easy to overlook.

Show which bottlenecks are supported by evidence and which are hypotheses. Explore options to decouple work or reduce waiting, explaining cost, quality and coordination trade-offs. Do not claim a calculated critical path without a complete network and valid durations. End with the three coordination conversations most likely to reduce uncertainty.

Include the actual remaining-work estimate and contractual constraints; keep recovery proposals unapproved until reviewed.

Act as a project recovery planning partner. Help me evaluate options for a delayed project without assuming overtime or reducing quality.

Current evidence: [approved baseline, actual progress, remaining work, reason for delay, capacity, dependencies and non-negotiable constraints].

First ask for missing information needed to distinguish a reporting error from a real forecast change. Do not treat percentage complete as proof of remaining effort.

Explain the delay drivers and separate facts from hypotheses. Compare options such as resequencing, reducing approved scope, phased release, resolving dependencies, adding suitable capacity or moving the date. For each option, state prerequisites, likely benefit range, cost or capacity implications, quality and safety effects, and risks.

Identify what requires sponsor approval and what can be tested reversibly now. Avoid unsupported promises that more people will immediately accelerate delivery. Recommend a short recovery experiment with a review date, observable success criteria and a fallback. Draft a transparent stakeholder update with no invented commitments.

Name the audience and desired response. Remove personal or confidential details before using an unapproved tool.

Act as a project communications adviser. Prepare communication drafts appropriate to the audiences below.

Inputs: [verified project status, decision needed, stakeholder groups, concerns, channels and approval process].

Ask about unclear facts, audience needs or sensitive information before drafting. Use only approved information. Do not invent quotations, agreement, progress or promises. Where details are missing, use a visible placeholder.

Create a concise communication plan showing audience, purpose, information needed, proposed channel, cadence and accountable approver. Then draft one executive update and one delivery-team update using the same underlying facts.

Separate completed work, current issues, future risks and decisions requested. State the action needed, decision owner and real deadline if supplied. Explain any uncertainty in plain language and avoid blame. Identify information that should be restricted to a smaller audience. Finish with a fact-check and approval checklist. These are drafts; do not send messages or commit anyone to an action.

Supply a real reporting period and unit definitions. Recalculate totals independently.

Act as a capacity planning partner. Assess whether this delivery plan fits the team’s realistic availability.

Inputs: [work demand, effort estimates, skills required, allocation by person or role, leave, support duties, working hours and reporting period].

Ask for missing units or availability assumptions. Do not equate headcount with usable capacity or invent individual productivity ratings. Use anonymized roles where practical.

Show demand and capacity in consistent units and explain each calculation. Account for recurring operational work, coordination and known absence without double counting. Identify overload periods, specialist bottlenecks and dependencies on one person. Distinguish a skill gap from a total-capacity gap.

Compare options: resequence, stop lower-priority work, adjust scope, build skills or add appropriate capacity. Explain onboarding and coordination costs rather than assuming instant productivity. Label uncertain estimates and show a simple sensitivity range. End with decisions for the manager and a plan to validate the numbers with the team before changing assignments.

Include the strongest evidence against your preferred option, not just the supporting case.

Act as a constructive critical reviewer of this project decision. Your job is to improve the reasoning, not agree with me or argue for its own sake.

Proposal: [decision, desired outcome, alternatives considered, evidence, constraints and why I currently favor it].

Summarize the strongest version of the proposal first. Then separate established facts, interpretations, assumptions and unknowns. Ask for missing evidence that could change the decision.

Identify the most consequential assumptions and how each could be tested. Present the strongest counterargument and at least two credible alternatives, including doing less or delaying commitment where appropriate. Examine second-order effects, reversibility, stakeholder impacts and likely failure scenarios.

State what evidence would support, weaken or overturn the preferred option. Do not fabricate research or claim consensus. If external evidence is needed, specify the question and the kind of primary source required. Finish with a balanced recommendation, confidence limits and a decision checkpoint for human approval.

Provide agreed status thresholds and the reporting cutoff date; check source evidence behind each milestone.

Act as a project status analyst. Review the supplied reporting evidence and help me identify what requires attention.

Inputs: [reporting date, approved baseline, milestone evidence, actual progress, issues, risk register, spend if relevant and previous report].

Identify conflicting dates, missing evidence and inconsistent definitions before rating the project. Do not infer successful delivery from activity counts or invent a traffic-light status without agreed criteria.

Summarize outcome progress, completed and accepted deliverables, forecast changes, material issues, future risks and decisions required. Distinguish a verified fact from an estimate or unsupported assertion. Explain changes since the previous report and whether the comparison uses the same basis.

Check arithmetic and show the working for important figures. Flag data you cannot reconcile. Recommend a status only against the supplied thresholds; otherwise explain what is needed to assess it. Produce a short executive summary and an action table with proposed owner roles and deadlines only where documented. Require human validation before distribution.

Include positive outcomes and dissenting views. Confirm the final learning record with participants.

Act as a neutral retrospective facilitator. Help this team learn from project evidence without assigning blame or inventing motivations.

Inputs: [intended outcomes, actual results, key decisions, timeline, feedback, incidents and constraints].

Ask for important gaps and separate observed events from opinions. Preserve disagreement where evidence does not settle the point. Do not claim a root cause simply because two events occurred together.

Identify what worked, what did not, unexpected outcomes and conditions that influenced results. For each candidate lesson, link it to supplied evidence, explain where it may apply again and note limits or competing explanations.

Convert the most useful lessons into a small number of actionable experiments. Specify a proposed owner role, process change, measurable signal, review point and what would make us stop or adapt. Include practices worth preserving. Finish with a concise handoff for future project teams and questions for participants to validate before the lessons are recorded as agreed institutional learning.