Evidence the run may read
- Customer request
- Transaction and service evidence
- Refund and exception policy
production blueprint · customer support & success
Refund request review agent blueprint.
Compare a refund request with policy, transaction, usage, and prior-case evidence and prepare a decision packet.
Search intent: AI agent for refund request review
Published by a2a cloudProduct-source methodology →
01 · the contract
Start from a bounded job, not a blank chat box.
trigger
A customer requests a refund or credit.
owner
customer experience owner
agent stops at
An authorized support or finance owner approves any refund, credit, or denial.
KPI · decrease
refund review cycle time (hours). Measure complete request to authorized decision.
outputs
Artifacts the run must produce
- Eligibility assessment
- Evidence and exception table
- Decision and response draft
02 · topology
Small specialists. Named handoffs. One accountable decision.
Each stage produces an artifact another stage can inspect. The final node is a person, not an autonomous write to an external system.
- 01
intake
Validate and normalize the refund request review inputs.
- 02
request normalizer
Produce the eligibility assessment.
- 03
policy evidence analyst
Produce the evidence and exception table.
- 04
refund approver reviewer
Challenge the refund request review result and prepare an approval packet.
- 05
customer experience owner
An authorized support or finance owner approves any refund, credit, or denial.
intake→ validated input packet →request-normalizer
request-normalizer→ Eligibility assessment →policy-evidence-analyst
policy-evidence-analyst→ Evidence and exception table →refund-approver-reviewer
refund-approver-reviewer→ approval packet with evidence references →human-approver
03 · authority
Grant the run only what this case needs.
Source material is read-only. Drafts land in a case-specific output path. Tools may read or propose; the human gate owns the external write.
case inputs
workspace/customer-success/refund-request-review/inputs/**Read only the evidence attached to this workflow instance.
case outputs
workspace/customer-success/refund-request-review/outputs/**Write drafts and evidence artifacts without modifying source records.
approved tools
customer-success:refund-request-review:read-or-proposeInvoke only tools explicitly granted for this run; external writes remain gated.
04 · implementation
A private, bounded starting manifest.
The blueprint starts private, caps its DAG, disables replanning, and exposes no public endpoint. Add only the tools and data adapters this workflow has approved.
a2a.yamlsafe starting point
name: customer-success-refund-request-review
version: 0.1.0
entrypoint: agent:BlueprintAgent
expose:
public: false
composition:
planning: deterministic_dag
max_nodes: 6
max_parallel: 1
max_replans: 005 · acceptance test
Pass only with evidence.
- 01
Evidence traceability
Every material conclusion cites an input artifact or a scoped tool result from this run.
- 02
The recommendation cites policy and transaction evidence and does not initiate a financial action.
The recommendation cites policy and transaction evidence and does not initiate a financial action.
- 03
Approval boundary
The run stops at a proposal and records the human decision before any external side effect.
failure containment
Stop small. Preserve the evidence.
Payment, service, or exception evidence is missing or contradictory.
Containment: Return a partial result with unresolved items; do not broaden scope or perform an external write.
Operator: Attach the missing evidence, narrow the brief, or explicitly approve a new scoped run.
A required input or tool grant is unavailable.
Containment: Stop the affected branch and preserve completed artifacts in the case output workspace.
Operator: Grant only the missing resource or continue with that branch marked out of scope.
06 · proof
Sign the run facts. Keep money in the billing ledger.
Current platform receipts sign caller identity or classification, skill, bounded input evidence, verified grant IDs when present, outcome or result preview, and timing. Optional file, tool, artifact, handoff, evaluation, and review fields require separate instrumentation and are not populated by default. Price, fees, payouts, and later human approvals remain separate platform records.
The example uses only fields populated by the current platform sealing paths. It is illustrative, not a record of a real customer run.
ExecutionReceipt · selected fieldsEd25519 token
{
"receipt_id": "rcpt_01J...",
"schema_version": 1,
"agent_name": "customer-success-refund-request-review",
"caller": "user:workflow-owner",
"task_id": "case_refund_request_review",
"skill_name": "refund_request_review",
"input_hash": "4d7c...9a2f",
"grant_ids": [
"grt_case_inputs",
"grt_tool_propose"
],
"status": "ok",
"result_preview": "Output prepared: Eligibility assessment. Human decision remains separate.",
"elapsed_ms": 4218
}related blueprints
Keep the operating model consistent.
Customer Support & Success
Customer support ticket triage
Classify a customer issue, identify missing context, and route it with an evidence-backed urgency proposal.
Customer Support & Success
Support response draft
Draft a clear answer from approved knowledge and case evidence, with uncertainty and escalation stated explicitly.
Customer Support & Success
Customer escalation summary
Consolidate case history, impact, commitments, and unresolved decisions into an executive-ready escalation packet.
put the pattern to work
Start private. Scope the authority. Require the decision.
Deploy the workflow as a bounded internal agent, verify its outputs and the receipt fields actually emitted, then expand only the scopes your acceptance test proves it needs.