AI Training Sessions

Your team uses AI tools informally and nobody can say what is safe to paste into them

Developers are integrating language models by trial and error

Leadership is being asked to approve AI spend without a way to judge the proposals

Training aimed at everybody at once

Developers, product managers and leadership need different material. A mixed room produces a session that is too shallow for the builders and too technical for the decision makers, so we aim each session at one audience.

Skills that decay in a fortnight

A workshop with no artefact leaves nothing behind. Every session ends with something the team owns, whether that is a working prompt library, an integration pattern or a written policy on what data can be used.

Tooling chosen for the demo

Sessions built around whatever is impressive this month age badly. We teach the parts that transfer, which is how to evaluate output, control cost and decide what should not be automated at all.

Developers get integration: how to call a model from your stack, structure outputs, evaluate them, control cost per request and decide what should never be automated. Product and operations people get judgement: how to spot a task worth automating, write a scope, set an accuracy threshold and read a vendor quote. Leadership gets decisions: what to budget, what the risks actually are under UK GDPR and the sector regulator, and which three rules to put in the staff policy this week. A mixed room produces a session too shallow for the builders and too technical for the decision-makers, so we do not run one.

Before the day, a short skills assessment and a look at the work your team actually does, so the exercises use your documents, your inbox and your systems rather than examples from a slide. On the day, hands-on from the first hour: every participant builds something on the tools the firm already pays for, ChatGPT, Claude or Copilot, with the firm’s own templates. Every session ends with an artefact the team owns, a prompt library, an integration pattern, a shared Project or custom GPT, or a written policy on what data may be used. A recording and the materials stay with you.

Half-day sessions cover one team and two tasks. Full days cover a team properly or two teams briefly. Multi-day programmes are for firms rolling AI out across the business and usually pair with the policy and the workspace setup.

Fixed prices, excluding VAT, delivered within three weeks of booking.

A business plan on ChatGPT or Claude costs a few pounds a seat. The waste is in seats bought and unused: firms that send a message and a login get a third of the seats used after a month, and the staff who do use AI drift to personal accounts for anything hard. Firms that adopt a one-page policy, build shared Projects on their own templates and run one session per team get most of the seats used and keep client data inside the sanctioned tool. In the 2026 recruitment survey we cite elsewhere, agencies that trained people reported far stronger returns than those that did not, and that pattern is not specific to recruitment.

Every session for a non-technical audience includes the data conversation, because it is the question staff most want answered and the one most policies leave vague. The short version: business plans do not train on your content and are covered by a data processing addendum; personal accounts on both ChatGPT and Claude use conversations for training unless the individual changes a setting; and special-category data belongs on an Enterprise plan or an in-region deployment with a DPIA, or nowhere. Leadership sessions walk through the ICO’s expectations in order and leave with the policy adopted.

How much do AI training workshops cost?

From £1,200 for a half day and £2,000 for a full day, with multi day programmes up to £6,000. The price covers curriculum design around your actual work, live delivery and the post workshop materials your team keeps.

Who are the sessions for?

Three audiences, each with different content: developers integrating language models, product managers deciding what is worth building, and leadership teams making budget and risk decisions. Mixed groups work, but the material is sharper when the session is aimed at one of them.

Is the training generic or tailored to us?

It is built around your workflows. Before delivery we run a short skills assessment and shape the curriculum around the work your team actually does, so the exercises use your processes rather than a sample dataset nobody recognises.

How soon can a session run?

Within three weeks of booking. That window covers the skills assessment, curriculum design and scheduling around your team rather than any queue on our side.