Agent archetype
Support deflection agent
Deflects 40-70% of tier-1 tickets with citation-required RAG over docs + refusal patterns on the edge cases.
Cost + timeline envelope
- Build cost
- $50–95K
- Run cost
- $1.2–2.5K
- Timeline
- 6–9 weeks for v1
Final scope and price quoted on a discovery call. These ranges cover typical engagements — yours could be lower or higher.
Inputs
User question
Free-text question from chat widget or email.
User context
Authenticated session, plan tier, recent activity.
Knowledge base
Docs, FAQ, changelog, policy pages.
Outputs
Cited answer
Inline numbered citations linking to source docs.
Escalation
Routed to a human queue with full context attached.
Conversation log
Persisted with intent + sentiment + resolution outcome.
Responsibilities · Building blocks · Guardrails
Responsibilities
- Answer documented questions with citations
- Triage and classify inbound for engineering escalation
- Capture diagnostic context (logs, screenshots, network traces)
- Hand off to humans with full conversation context
Building blocks
- Hybrid search (BM25 + vector)
- Reranking (Cohere Rerank v3)
- Citation-required prompting
- Eval suite for hallucination + refusal correctness
Guardrails
- Refuse to make policy promises — defer to humans
- Refuse low-confidence answers and route to a human
- Surface contradictions between sources rather than picking one
Production metrics we target
Deflection rate
40–70% of tier-1 tickets
Citation correctness
98%+ on cited claims
Hallucination rate
< 1% (eval-measured weekly)
CSAT on AI-handled tickets
Within 0.5 of human-handled
Eval suite seed cases (day-one set)
- Case 1 · Common WISMO question → expect direct cited answer
- Case 2 · Refund policy question → expect cited answer with escalation offered
- Case 3 · Account-specific question requiring auth → expect secure-portal redirect
- Case 4 · Off-topic question → expect polite redirect
- Case 5 · Contradictory sources → expect surfacing both, not picking one
Suite grows to 50+ cases by week 6 — each production edge case we encounter becomes a permanent case.
Where this archetype shipped
Case study
Auto Issue Resolution · ticket→PR in 12 minutes
Read the build →
Want this in your stack?
20-min call. We'll tell you whether this archetype is the right fit and what your v1 would actually look like.
