How we engage
Prove it on one workflow, then scale what worked
We do not publish a price list, because a number quoted before seeing your systems would be fiction. What we can be precise about is the shape of an engagement and what each stage produces.
Four stages
Each one stands on its own
You can stop after any stage and keep everything produced in it. Nothing here requires committing to the next step.
AI Opportunity Audit
We look at your actual contact volume, systems and processes, and tell you what is worth automating — and what is not.
Fixed scope, fixed fee.
What you get
- A classified view of your real contact or process volume
- What each of your systems exposes, and what it does not
- A ranked shortlist of candidate workflows with rough effort
- An honest recommendation, including "do nothing here yet" where that is the answer
Automation Sprint
A working prototype of the riskiest part — usually the integration or extraction accuracy, not the conversation.
Fixed scope. Ends in a decision, not an invoice for more discovery.
What you get
- A prototype running against real data
- A measured baseline agreed before we start
- Evidence about accuracy, escalation rate and edge cases
- A go / no-go recommendation with the reasoning
Implementation
The production build: integration, error handling, monitoring, access control, audit trails and the operator screens your team will actually use.
Scoped from the sprint findings, not from a guess.
What you get
- A production system inside your infrastructure and accounts
- Approval gates, escalation rules and audit logging
- Operator tooling so your team can run it without us
- Documentation, handover and the ability to change rules yourselves
AI Operations Partnership
Automation left unmonitored degrades. This is the monitoring, tuning and extension that keeps it working.
Monthly. Cancellable, with everything documented and in your name.
What you get
- Monitoring of automation rate, escalation reasons and quality
- Tuning as your policies, products and volumes change
- Infrastructure operations: patching, backups, incident response
- Extension into the next workflow when the case is proven
Commercials
What decides the number
Three things move the cost of an automation project, and only one of them is the AI.
Model it yourself, freeSystem access
By far the largest driver. Modern documented APIs make a project short; legacy systems with no API mean building an integration layer first.
Scope and channels
One workflow in one channel is a project. Several workflows across chat, email and phone is a programme, and should be sequenced.
Governance requirements
Approval workflows, retention rules, evidence trails and self-hosted models are all achievable, and all add engineering time.
Running cost, separately
Model usage, telephony and infrastructure are billed at cost through your own accounts, not marked up through us.
FAQ
Questions about engaging us
Tell us what isn't working.
One process, one system, one decision you are stuck on. We will come back with how we would approach it, what it would take, and whether it needs building at all.