Custom AI Agent Design Brief

Act as an AI systems architect. Write a build-ready specification for a custom AI agent.

What it should do: [TASK DESCRIPTION] · business function [e.g., customer support, lead qualification, invoicing] · task volume [times per day/week] · who does this now [role, time spent] · perfect result [describe] · must never happen [describe failures] · needs access to [tools/systems] · who interacts with it [customers / team / both]

Design:
1. Production-ready system prompt — persona, capabilities, constraints, tone, escalation rules
2. Knowledge base requirements — documents, data, FAQs
3. Tool integrations — APIs with read/write/trigger specs
4. Guardrails — hard limits, escalation triggers, confidence thresholds
5. The 5 most common conversation flows as decision trees
6. 10 test cases (5 normal, 3 edge, 2 adversarial) with expected outputs
7. Success metrics — resolution rate, time saved, escalation rate, satisfaction

Success means every "must never happen" item has a matching guardrail and at least one test case, and the system prompt states each rule once.

Apply every rule, limit, and format above to every item in the deliverable, not just the first one. Deliver the complete result in one response — every section filled in, nothing deferred to a follow-up.

If a required input is missing, ask for it once before starting; otherwise proceed and label any assumption you made.

Notes: R40 refresh; tuned to vendor prompting guidance