9. User Journeys
9.1 Journey A: Independent SaaS builds a Customer Support Copilot
Brief: Triage tickets, draft replies, cite knowledge base.
Context Pack: Help center, resolved tickets, product docs, glossary.
Agent: Triage Agent → Research Agent → Draft Agent → Risk Gate Agent.
Eval: 200 golden tickets; accuracy, grounding, tone, escalation.
Release: Master Agent v1.0 → Tokenized; 92% first-contact resolution in pilot.
9.2 Journey B: Auto Dealer Group (Dryve-style) creates Multi-Agent Sales Network
Brief: Model matchmaker, valuation, negotiation, DVLA-style transfer helper.
Context: Listings, pricing history, policy docs, DVLA API stubs.
Agents: Valuation, Negotiation, Compliance, Orchestrator.
Eval: Offer fairness, compliance coverage, citation accuracy.
Release: Master Agent v1.0; 18% faster conversion, 12% higher blended margin.
9.3 Journey C: Restaurant Group (Plates AIA-style) deploys Ops Copilots
Brief: Daily briefings, menu engineering, compliance, vendor swap suggestions.
Context: SOPs, recipes, allergens, invoices, vendor SKUs.
Agents: Pepper (Ops), Plato (Menu), Vendr (Procurement), Hawk (Compliance).
Eval: Food cost variance, allergen accuracy, suggestion acceptance.
Release: Master Agent Suite; 9% food cost reduction, 22% faster onboarding.
Last updated