Kick-off
Goals, success metrics and the stakeholder map, agreed with leadership.
Five phases. Each one leaves you with something you own, even if you stop there. The order matters: we never automate a process we haven’t mapped, and we never roll out a brain to a team that hasn’t learned to work with AI.
Goals, success metrics and the stakeholder map, agreed with leadership.
6–7 one-to-one interviews in the most willing departments — usually marketing and HR. We understand the processes and, at the same time, bring the team into the change.
Sessions with managers and key people in the other departments, once the pilot has shown its value.
Current processes, tools and manual load by department; a readiness survey and score; a list of automation candidates ranked by effect and effort.
A presentation-ready plan for leadership with phases and expected impact, plus 3–4 quick wins ready to start.
Audit page: AI readiness audit · £4,500 · 2 weeks.
Hands-on sessions built on your real tasks and documents: prompting, ChatGPT and Claude as workspaces, a knowledge assistant on your own files, presentations and data, agents and automation with n8n and MCP, AI safety. Every session is recorded.
Training is also discovery. We see the processes from the inside, and people stop resisting the change because they get a personal benefit first. A before-and-after test shows leadership the change in numbers.
Training pages: AI training for teams · AI training for leaders · £950 per session.
Call notes into tasks and a channel summary. Proposals from call transcripts with your pricing logic. Weekly reporting from the data you already have. Candidate profiles from an application form.
Each pilot takes 1–4 weeks, goes live in production and is measured before and after. If a process is deterministic, we automate it without AI — a language model only adds cost and unpredictability where a plain integration will do.
Automation page: AI automation · from £4,000 per process.
An AI use policy, a shared prompt library and simple rules your team will actually follow: what can and cannot go into a model, how to check an answer, who owns which automation.
Rules written apart from the work get ignored. We write them while the pilots are running, from what the team is actually doing.
We connect your sources, build the brain and roll it out in Slack or Telegram. From 2 weeks on ready-made tools to about 2 months for a custom build with its own knowledge graph.
The working brain on your infrastructure if you choose, the data pipelines, the documentation — and the option of ongoing support from £3,500 a month.
Product page: Company brain · from £6,000.
No. Each phase leaves you with something you own, and you can stop after any of them. Most clients start with the audit or with training and decide on the rest with the roadmap in hand.
The audit takes 2 weeks. Training runs 3–6 weeks depending on the format. Each automation goes live in 1–4 weeks. A company brain takes from 2 weeks on ready-made tools to about 2 months as a custom build.
One owner for the project, plus 6–7 people for interviews during the audit and the team members who join training. Without an owner on the client side, new tools don’t stick — we’ve seen it, so we ask for one.
Before anything is built we record the current numbers: time per task, people using AI, time spent searching. Every pilot and the brain are measured against them afterwards. You get the before-and-after figures in writing.
Most companies start with the 2-week audit. Book a call and we’ll tell you honestly whether that is the right first step for you.