AI Agent and Chatbot Development

The failure mode is consistent. A generic chatbot goes live, answers plausibly for a week, then tells a customer something that is not true — a policy that does not exist, a price that is wrong, a commitment nobody can honour. The problem is not the model; it is that nothing constrained what it was allowed to claim.

What we build

• Customer-facing assistants over your documentation, policies and product information, with escalation to a person.
• Internal assistants so staff can retrieve an answer instead of asking a colleague.
• Agents that take actions in your systems — raising a ticket, updating a record — with permissions and logging.
• RAG pipelines — ingestion, chunking, retrieval and citation over your own content.
• Evaluation harnesses, so a prompt or model change can be measured rather than guessed at.

Chatbot or agent?

Frequently asked questions

How much does an AI chatbot cost in the UK?

A grounded assistant over existing documentation typically £10,000–£30,000 to build, plus running costs per conversation. Agents that take actions cost more, because permissions, confirmation flows and audit logging are most of the work.

Will it make things up?

Less, with grounding, citation, tight scope and an explicit refusal path — but no honest supplier will tell you never. That is exactly why customer-facing assistants need an escalation route and why anything consequential stays with a person.

Can we use our own data without it training a model?

Yes. Retrieval does not train anything — your content is fetched at query time and passed as context. Enterprise API tiers generally do not train on submitted data either, and where data cannot leave your environment at all, we design for that constraint from the start.

How do we know it is working?

An evaluation set of real questions with known-good answers, run on every change, plus the business measure — deflected tickets, time saved, escalation rate. If nobody can name that measure, the assistant is not ready to be built yet.

How the engagement runs

• Scoping call, then discovery. We establish what you actually need before quoting. You approve scope before we write a line of code.
• Fixed price or capped time and materials. Tightly specified work suits a fixed price; genuinely new work suits a capped retainer. We recommend one rather than offering both and letting you guess.
• Weekly demos. You see working software every week, so nothing surprising turns up at the end.
• You own everything. Repositories, infrastructure and documentation transfer to you on payment. No vendor lock-in, no licensing you do not control.
• NDA as standard, usually before the first detailed call. We will often send one ourselves if you do not.

If you are planning work of this kind and want a straight assessment of scope, timeline and cost, tell us what you are building. We work with clients across the UK, EU and US.