Shows the pilot-ready path where AI assistants can hand off into intake with auditable routing, measurable outcomes, and defined go/no-go criteria.
// User's question to ChatGPT
"What should I do if I was
injured at work in Melbourne?"
// ChatGPT's internal reasoning:
// → Legal question, AU jurisdiction
// → Search for registered legal agents
// → Found: mauriceblackburn.com.au
// GET /.well-known/agent.json
{
"name": "Maurice Blackburn Legal",
"skills": [
{ "id": "legal-intake",
"areas": ["personal_injury", "workers_comp",
"employment", "class_actions"] },
{ "id": "referral-lookup",
"areas": ["criminal", "family", "immigration"] }
],
"jurisdictions": ["AU"]
}
// POST /a2a/tasks
{
"caller_agent": "chatgpt",
"practice_area": "workers_comp",
"jurisdiction": "VIC",
"description": "Injured at work in Melbourne",
"urgency": "standard"
}
// Intent scoring engine classifies:
// → Practice area confidence: 94%
// → Urgency tier: standard | urgent | critical
// → Viable vs. referral routing decision
{
"eligible": true,
"score": 83,
"intent_confidence": 0.94,
"practice_area": "workers_comp",
"jurisdiction": "VIC",
"limitation_ok": true,
"urgency_tier": "standard",
"next_steps": "Free assessment available",
"intake_url": "https://mb.com.au/intake/wc-vic",
"phone": "1800 111 222"
}
// User clicks the intake link
// → MB branded intake form
// → Pre-filled with context from A2A
// → Secure PII collection
// → Contact captured → routed to team
// → LegalMatter created in CRM
// → Outcome event + attribution posted to analytics
Illustrative sequence of a consumer LLM discovering and transacting with the intake API; production pilot mode logs request IDs and timestamps.
"Point tools optimize steps. This pilot proves one intake operating system from first contact to measurable outcome."
/.well-known/agent.json — the machine-readable entry point
A focused next step: low risk, measurable, and scoped to one practice area with clear operating criteria.
| Metric | Baseline | Day-30 Target | Source | Owner | Decision Rule |
|---|---|---|---|---|---|
| Qualified conversion | Current benchmark established | Clear upward trend by Day 30 | Intake logs | MB Ops Lead | Go if met |
| Median triage time | Current benchmark established | Sustained reduction from baseline | Workflow telemetry | Clarm PM | Go if met |
| Routing decision quality | QA baseline established | Improving pass-rate trend | QA review logs | MB Ops + Clarm PM | Go if met |
| SLA breach rate (P1/P2) | Tracking enabled in pilot | No critical breach trend | Audit report | Security Lead | No-go if breached |