Use this skill to convert partner SPIF dispute, channel incentive dispute, referral payout conflict, reseller rebate dispute, deal registration conflict, attribution claim, payout calculation, clawback, and partner appeal questions into a concrete artifact with owners, gates, metrics, and recovery paths.
Workflow
- Identify program terms, partner claim, eligible action, deal/account, attribution source, payout formula, evidence packet, approval authority, dispute SLA, fraud indicators, clawback state, communication owner, and learning path.
- Read
references/partner-spiff-dispute-resolution-patterns.md. - Classify the situation as eligibility dispute, duplicate attribution claim, payout calculation dispute, late deal registration, clawback dispute, partner tier dispute, fraud suspicion, or appeal after denial.
- Define dispute intake, evidence requirements, source-of-truth order, reviewer/approver roles, calculation audit, partner communication, SLA, appeal route, clawback handling, fraud escalation, and program feedback.
- Produce SPIF dispute resolution review, state machine, decision table, event schema, evidence checklist, partner communication plan, and incentive-program learning loop.
When not to use
- Do not use for generic advice the base model already handles without this skill's specific artifact contract.
Guardrails
- Do not resolve disputes from relationship pressure instead of program terms, evidence, and independent approval.
- Do not let sales notes or spreadsheets override the incentive source of truth without exception approval.
- Do not surprise partners with undisclosed eligibility, clawback, tax, invoicing, or timing rules during a dispute.
- Do not treat disputes as one-off noise; feed patterns back into program terms, attribution, partner education, and fraud controls.
Output format
Partner SPIF dispute context:
Audience / source of truth / risk boundary:
Dispute resolution plan:
| Area | Decision | Evidence | Risk | Owner |
| --- | --- | --- | --- | --- |
Eligibility, attribution, evidence, calculation, approval, communication, appeal, clawback, and program learning:
- <trigger> -> <policy, metric, edge case, support note>