Three ways to work with us. No recurring "AI seats," no per-conversation pricing, no platform lock-in. You pay us to build something, then you own it.
Quoted in GBP, excluding VAT. Larger or longer engagements: let's talk.
Support, account questions, order tracking. Deployed to web, WhatsApp, Messenger, Telegram, voice, or your mobile apps.
Domain-trained agents connected to your CRM, ticketing, knowledge base, and APIs. Audit trail and RBAC included.
The unglamorous middle layer: telephony, payments, identity, evals. We make assistants work inside your real stack.
Tracks parcels, reschedules deliveries, changes addresses, handles shipping FAQs. Built in a 6-week production engagement.
Account-aware support, onboarding, scoped tool calls, PII scrubbing, audit log. Built in an 8-week engagement.
Hourly billing penalises us for working efficiently and you for being uncertain. A fixed scope and fixed fee make us both honest about what's being delivered.
Third-party costs: model API spend (OpenAI, Anthropic, etc), hosting (AWS, GCP), and any per-conversation fees for channels like WhatsApp Business. These are billed to you directly, not through us. Typical model spend for a customer support assistant is £50-£500/month.
A single-channel assistant with one tool is £18k. A multi-channel assistant with 6 integrations, evals, audit, and a review queue is £65k. The Discovery week gives us both a real number, not a range.
Yes. It lives in your repository under your licence from day one. You can fork us out tomorrow and the assistant keeps running. No vendor lock-in, no per-seat fees, no recurring "AI subscription."
Yes. We sign DPAs, work inside your VPC, and use private model endpoints where the data demands it. We've built for banking and we're comfortable with the constraints.
Short brief is fine. We reply within one working day and quote within two.