AI agent team guide
AI Capacity Planning: Resolve Conflicts Across Projects
Reconcile competing commitments across projects and prepare explicit options before a human accepts new work.

Use AI capacity planning to make a tradeoff visible before a team silently accepts too much work. This workflow reconciles commitments across projects, identifies the actual constraint and prepares options for the person who can choose what changes. It does not turn a calendar estimate into guaranteed delivery.
Build your four-agent team
| Agent template | Its job on this team | Give it | Reviewable return | Team |
|---|---|---|---|---|
Capacity PlannerLead agent | Reconcile available capacity and competing commitments. | Owner-supplied estimates and the shared planning horizon. | A constraint map and tradeoff options. | Add to team |
Project ManagerSupporting agent | Explain project milestones and dependencies. | Accepted project plans and current status. | Which commitments move if an option is chosen. | Add to team |
Executive AssistantSupporting agent | Check the decision calendar and coordination needs. | Review dates, availability and accepted owner list. | Scheduling questions and a decision-ready agenda. | Add to team |
Decision Journal KeeperSupporting agent | Record the human tradeoff and revisit trigger. | Accepted option, owner and assumptions. | A commitment record with explicit deferrals. | Add to team |
Set up your team in Finta
- Add each coworker individually. Use each Add to team action, review the import setup and complete that template’s Getting Started Skill. You can also open its marketplace profile to review its instructions and requirements first.
- Supply role-specific context. Give the lead the shared objective. Give each supporting agent only the authorized evidence and accepted prior return it needs. Importing the templates does not automatically connect outside systems or share all context.
- Review each handoff. Accept or correct the return before passing its evidence and open questions to the next role. Ask Aurora to help coordinate the work while you retain the decisions.
- Start with one manual pass. Review any supported routine separately before enabling recurrence. External messages, record changes and commitments need their own authorization.
Prepare your inputs
Bring the same planning horizon for every project, committed milestones, routine work, known leave, dependencies and owner-supplied effort estimates. Use comparable units. Ask people to validate their availability; a full calendar is neither a complete workload record nor permission to infer individual productivity.
Best practices worth bringing into the work
Atlassian’s capacity-planning play includes routine work and competing commitments. Its capacity guide treats the plan as something to revisit. Apply both by counting the work already promised and recording what must move when something new is accepted.
Run the workflow
1. Normalize the planning window
Capacity Planner puts the supplied estimates on the same period and unit. Keep uncertain estimates as ranges. Separate routine work from project effort so support, reviews and maintenance do not disappear from the calculation.
2. Find the shared bottleneck
Project Manager maps milestones to the constrained skill, person or external dependency. Team-wide unused hours do not solve a specialist bottleneck automatically. Have the responsible people validate the map.
3. Prepare real alternatives
Compare deferring a request, reducing scope, changing sequence or adding authorized capacity. Name the effect on existing promises. Executive Assistant organizes the decision discussion around the unresolved tradeoff rather than another general status report.
4. Accept the change explicitly
Decision Journal Keeper records the chosen option, what was not accepted and the event that triggers another review. Update actual commitments only through the authorized owner. Keep the old assumption visible when new work changes the plan.
Cross-project commitment and tradeoff board
Use this original worksheet with your own authorized evidence. The entries below are prompts for a review, not completed work.
| Commitment | Constraint / evidence | Option | Decision owner |
|---|---|---|---|
| Project A release | Requires the same review owner as B. | Sequence the reviews or reduce scope. | Project owners accept the tradeoff. |
| Project B pilot | External dependency not confirmed. | Hold the dependent commitment. | Pilot owner confirms readiness. |
| Routine support | Known recurring work in the same window. | Reserve capacity before adding projects. | Operating owner validates the assumption. |
Worked example
Fictional arithmetic: one specialist has 30 hours of available work time in the chosen window. Routine work requires 10 hours, Project A estimates 14 and Project B estimates 12. That is 36 hours of commitments against 30 available. The decision is what six hours or equivalent scope will move, not how to label both projects “on track.”
Copy the team brief
Review capacity across [projects] for [shared horizon]. Use owner-supplied estimates and include routine work and known leave. Identify shared bottlenecks, uncertain assumptions and dependent milestones. Compare defer, reduce-scope and resequence options. Return the decision needed and the commitments affected; do not promise dates or change assignments.Review the return
An accepted plan names the chosen tradeoff, affected promises, responsible owners and next review trigger. If the estimates cannot be normalized or a key owner has not confirmed availability, publish that uncertainty inside the team’s plan rather than a precise but unsupported capacity number.
Before you start
Why can a team have spare capacity and still miss a milestone?
Capacity is not always interchangeable. A particular review, skill or external dependency may constrain several projects even while other work time is available. Plan against the real dependency, not only a team-wide total.