AI Chatbots & Knowledge Assistants
Trained on your own content — docs, pricing, policies. Accurate answers, 24/7. Not a decision tree, not a GPT wrapper.
from £1,500Open for new projects
Most businesses don't need a six-figure AI strategy. They need one well-built automation that saves the team eight hours a week — and a developer honest enough to say so.
The Honest Version
There's a lot of noise. Most of it is hype from consultancies selling roadmaps, or cheap tools that don't do what they claim. The honest version: AI is genuinely useful for a specific set of problems — repetitive tasks, predictable inputs, clear outputs. Applied there, it saves serious time and money. Bolted onto the wrong process, it wastes everyone's.
My job is to tell you the difference — and only build what's worth building.
What I Build
Not "AI strategy". Not roadmaps. If you can describe the problem, I can tell you whether AI fixes it and roughly what that costs.
Trained on your own content — docs, pricing, policies. Accurate answers, 24/7. Not a decision tree, not a GPT wrapper.
from £1,500Your document library, queryable in plain English — answers with source references, without the hallucinations.
from £2,000Engages visitors, asks the right qualifying questions, routes hot leads to you instantly. Never sleeps, never misses one.
from £1,200Self-hosted pipelines connecting your tools. Zapier-class power without the per-task tax — documented and handed over.
from £500 / workflowReal AI features inside your existing app — summarise, classify, extract, search. With proper error handling and cost controls.
from £800Contracts, invoices, and forms in; structured data out. Manual data entry, eliminated from the workflow.
from £1,500Inbound calls, reminders, out-of-hours capture. An emerging field — I'll be straight about what works yet and what doesn't.
from £2,500Two to three focused hours on your workflows, then a plain-English prioritised roadmap you can take to anyone.
£350–£600The Right Fit
AI integration isn't the right answer for everyone. I do my best work with:
Process
We start with the problem — your workflow, your data, what “good” looks like. The right tool emerges from that conversation. I don't arrive with a solution looking for a problem.
Fixed scope, fixed price: what gets built, what it connects to, what it costs, and a realistic timeline. AI projects carry more uncertainty than web builds — where it exists, I'll say so.
Most implementations have a working prototype within 1–2 weeks. We test against real data and tune the behaviour until it does what you actually need. You're involved throughout.
Full documentation, monitoring, and cost controls — no surprise API bills. Ongoing support on retainer or ad-hoc, your call.
The Self-Hosted Advantage
Most AI tools — Zapier, Make, ChatGPT plugins, chatbot SaaS — process your data on their servers. You pay per use, and your business data flows through third-party infrastructure. For many businesses that's fine. If you handle client confidentiality, legal privilege, or sensitive personal data — it isn't.
It costs more than the SaaS route. For law firms, healthcare, and financial services, it's the right call.
Tooling
The tools depend on the problem — there's no single AI stack that fits everything. This is the landscape I work across.
Investment
AI projects vary enormously by complexity, data volume, and integration depth. As a guide:
These are starting points — everything is quoted at a fixed price after a scoping conversation, no hourly billing surprises. Ongoing API costs (OpenAI, Anthropic, etc.) are separate and depend on usage; you'll get a realistic running-cost estimate before you commit.
Typically replies within one business day Remote delivery worldwide