AI training for teams is a hands-on workshop where staff learn to use tools such as ChatGPT, Claude, Copilot or Gemini on their own tasks, with clear rules on what data stays out. Praxen runs a half-day session for up to 12 people, online or on site (travel extra), with a follow-up two weeks later, for £850 (£595 with the launch offer). Prices exclude VAT.
Built on your team’s own work
Many people try ChatGPT at work with a blank box and a vague question. They get a generic answer, decide the tool is of little use and go back to the old way. A session built on the team’s own emails, documents and reports gives each person a working example for a task they do every week. That example is what keeps people using AI after the session ends.
The second reason is safety. Some staff may already paste work into AI tools on personal accounts. Training is the moment to agree which tools are approved, which accounts to use and what data never goes in.
The exercises depend on who is in the room. These are typical tasks we build them around:
- Sales: preparing for a call from the account history, follow-up emails and first drafts of proposals.
- Customer support: reply drafts from the help centre, summaries of long complaint threads and a consistent tone across the team.
- Finance: commentary on monthly figures, checks on supplier statements and plain-English summaries for managers.
- HR: job adverts, answers to policy questions with the handbook section quoted, and interview question sets.
- Operations: step-by-step procedures from rough notes, meeting minutes with actions and owners, and supplier comparisons.
Before the session
- A 30-minute call with the team lead about goals, roles and the tools the company allows.
- A short questionnaire for each participant: which tasks take the longest and how they use AI today, if at all.
- Five to ten sample documents, such as emails, reports or templates, with names and personal details removed.
- A check of accounts: everyone needs a login to the approved tool, on a business plan, before the day.
- A written programme for you to approve, about a week before the session.
The half-day programme
A typical agenda for a mixed team. We adjust each part to your roles and tools.
| Part | What people do | What they leave with |
|---|---|---|
| Rules and basics | How the tools work, where they get things wrong and what data stays out | A draft usage policy |
| Written work | Drafting customer emails, proposals and meeting notes from the team’s own examples | Prompts that match your tone |
| Documents and research | Summarising long documents, pulling out figures and checking answers against the source | A checking routine for every answer |
| Your own task | Each person builds and tests prompts for one task they do every week | One working prompt per person |
| Playbook and next steps | Good prompts go into the shared library, and everyone picks a task to use for two weeks | The team prompt library |
What your team keeps
- A prompt library for your own tasks, stored where the team already works, such as SharePoint, Google Drive or Notion.
- A written usage policy, usually one or two pages, ready for you to adopt.
- A list of approved tools and accounts, with the settings to check on each.
- Notes and worked examples from the session.
- A follow-up Q&A two weeks later, online, where we fix prompts that did not work and measure hours saved per person.
- A short list of the tasks people said take the most time. It often points to the first workflow worth automating, if you decide to go further.
What the usage policy covers
We draft the policy from your existing IT and data protection rules, so it fits what you already have. It stays short enough for people to read, and the team goes through it together in the first part of the session.
- Approved tools and accounts. Business and enterprise plans from the main providers do not train models on your data by default. Free personal accounts may, depending on their settings.
- Data that never goes in: client personal data, health information, passwords and unpublished financial results, unless the tool and the contract are approved for it.
- Checking: a person reads and checks every AI output before it is sent, filed or relied on.
- Disclosure: when to tell clients that AI helped with a piece of work.
- Ownership: who answers questions about the policy and who approves new tools.
Format: online or on site
Online sessions run on Microsoft Teams, Zoom or Google Meet. On-site sessions take place at your office, and travel is charged extra. Each session takes up to 12 people, so everyone gets time on their own task. Larger teams are split into groups, ideally by role, so the examples fit the work: one group for sales and support, another for finance and operations.
Each participant needs a laptop and access to the approved tool. People with no AI experience are welcome. The first part covers the basics, and each exercise starts from the person’s own task, so it works at any level.
If a team works with sensitive data, such as client files or health records, we build the exercises on anonymised copies and agree in advance which tool settings must be switched on.
What corporate AI training costs
One price covers the session, the prompt library, the usage policy and the follow-up. Prices exclude VAT. Licences for ChatGPT, Copilot or other tools are paid to the providers on your own accounts. A team of 30 would need three sessions.
The figures in this payback check are illustrative. If 12 people each save one hour a week and their time costs £25 an hour, that is £300 a week. On those figures, a session at £595 covers its cost within a month. Your figures will differ, and the follow-up is where we measure them.