A trade front-office agent for the hours your team cannot cover. Answers, quotes, books, and leaves evidence.
Tradie Front Office AI routes every inbound inquiry through one Cloudflare-native agent loop: approved business knowledge, deterministic pricing, D1 persistence, and an audit trail your team can inspect.
"I have a blocked toilet in Kedron and water backing up in the shower. Can someone come today and what will it cost?"
Retrieves warranty, service-area, and safety policy snippets.
Filters service windows by job type, suburb, urgency, and availability.
Runs deterministic rate card math before quoting.
Creates a job or hold with payment pending for demo.
The exact customer request and agent reply remain visible.
Staff can inspect availability, price, and job steps.
Low-certainty or policy-heavy requests route to staff.
The operational outcome appears in the dashboard.
High-intent customers arrive outside the staffing model.
Businesses lose revenue in the exact moments customers are ready to act: after-hours calls, weekend texts, email backlogs, and policy questions that stall a quote. Tradie Front Office turns those moments into traceable workflow.
The caller wanted a price and a service window now, not a callback tomorrow.
A job-ready question sits beside low-priority admin noise.
Staff need the agent to quote from the rate card, never from vibes.
Policies help the agent answer, but cannot override availability or pricing tools.
Voice, SMS, email, and web all land in the same operating surface.
The channel can change. The proof does not: memory, tools, persistence, staff review, and escalation all follow the same route.
Channel metadata, customer details, and transcript become one inquiry record.
Enabled Agent Memory chunks enter the prompt as governed context.
Availability, pricing, job, and escalation stay typed and inspectable.
Inquiries, jobs, and events are written to D1 with business_id scope.
Every channel appears in the same queue.
The why behind each answer is visible.
Confirmed holds and pending payments are clear.
Risky requests escalate instead of improvising.
The AI is useful because the business stays in control.
The demo is designed around trust: no invented prices, no invisible policy leaps, no mystery jobs. Staff can see the source, the tool, the confidence, and the outcome.
Pricing is deterministic
Every quote comes from calculatePrice and the quote_price tool.
Memory is bounded
Uploaded policies inform answers but cannot override pricing or availability.
Roadmap is labeled
Stripe, ServiceM8, R2, Vectorize, Queues, and AI Gateway stay roadmap until real.
Open the product and follow one request end to end.
Start at the dashboard, run a channel demo, inspect the inbox record, then open the tool trace and pricing proof. The homepage is the promise; the dashboard is the evidence.