AI Agent Knowledge
The knowledge sources that power the Qontak AI Agent. Today the AI Agent learns from static sources (uploaded PDFs and websites/URLs); this initiative extends that to dynamic, proprietary sources so clients can activate a high-quality knowledge base from data they already own — lowering the "cold start" barrier to AI adoption. Phase 1 adds Conversation History: learning from high-performing human agents' resolved conversations.
QA Lane
Lane B — keeps a human QA gate. The spec is still too thin to positively rule out Lane-B triggers, and no E2E test specs exist for this initiative yet, so the Lane-A entry bar (100% E2E, spec-mapped coverage) is unmet — Lane B by default. Classified 2026-06-29.
Master index (ANCHOR)
ai-agent-knowledge-anchor.md— the ANCHOR PRD: the initiative master index (identity, Source Index, north-star metrics, decisions). It carries no acceptance criteria of its own (those live in the phase PRDs), so it sits at the initiative root rather than underprds/. Synced with the Confluence AI Agent Knowledge: Conversation History and reconciled against code (chatbot,chatbot-fe,qontak-designer).
Phases (by knowledge source)
| Phase | Source | Status | Epic |
|---|---|---|---|
| Phase 1 — Conversation History | Resolved conversations of selected expert agents | PRD draft (Discovery) | TBD |
| — | PDF / Website (static sources) | Shipped (pre-existing) | — |
| Phase 2 — Data-driven agent selection | System-suggested experts (CSAT/AHT) instead of manual pick | Planned (future consideration) | TBD |
Static PDF/URL knowledge already ships in production; this initiative captures it as the umbrella and adds dynamic sources. Future phases (data-driven selection, multi-media ingestion, real-time sync, cross-division pooling) are seeded from the Phase 1 "Future Considerations" — placeholders until their phase PRDs are written.
Contents
prds/— phase PRDs (each with its own ACs → Jira Epic) land here.rfcs/— Request for Comments (technical design proposals).tests/— E2E / acceptance test specs.