Your team on Claude or ChatGPT, with one set of rules.
Business accounts in the company's name, your approved documents loaded, a shared set of prompts for the tasks you repeat, and one page that says what goes in and what never does. Then a session where everyone practises on their own work.
One page, three columns, taped next to the screen.
A sample. We write yours with you, using your own examples, so nobody has to guess.
Fine to use
- Your price list and published policies
- Email drafts with names removed
- Templates, product descriptions, FAQs
- Meeting notes you wrote yourself
Ask first
- A specific customer's order history
- Supplier contracts and quotes
- Anything you would not put in a group email
- New tools or plug-ins not on this sheet
Never paste
- Card numbers, bank details, passwords
- Health or patient information
- Anything into a personal AI account
- Staff records and pay
And always: a person reads it before it goes to a customer.
Before, during and after the session.
Each person lists five repeated tasks
With a real example of each: a quote reply, a supplier chase, a job summary. These become the practice material.
We set up the accounts
Business plan, company billing, two-step login, and projects loaded with the documents you approve. Nothing personal.
We write the prompts and the sheet with you
A shared library for the repeated tasks, in your tone of voice, and the rule sheet above with your own examples.
Hands-on session
Everyone does their own tasks on the new setup while we sit beside them. No slides about the history of AI.
A check a few weeks later
What is being used, what was abandoned and why. We fix the prompts or rules that did not stick.
Two founders, AI every day, the same rules.
We run MUVES with Claude, ChatGPT and Codex inside our own operating system. The habits we teach are the ones we keep:
ClaudeChatGPTCodex- Business accounts onlyand only the information a task needs
- Drafts wait for a personnone of our agents can send an email on their own
- Instructions reviewed firstbefore they reach a prompt, and kept short
That last one matters for teams too. A prompt library that grows without anyone checking it drifts. Ours has a reviewer and a size limit; yours gets an owner. See how MUVES OS works.
Your team, for a morning.
- Who needs accessand what each person does
- Five repeated tasks per personwith real examples
- The documents the team should work from
- A block of timewith everyone in the same room or call
Asked in every session.
Claude or ChatGPT?
Both are capable. We choose based on the tools your team already uses and the tasks on your list, and we explain the choice in plain words. We are not a partner or reseller of either company.
Will our data be used to train the models?
We show you where each vendor's business terms address training on your data and set the account's options accordingly. The rule sheet keeps the most sensitive information out regardless.
What if people go back to their old habits?
That is why the session uses each person's own repeated tasks. A few weeks later we look at what is being used and fix the prompts or rules that did not stick.
More for teams starting with AI
Rules, assistants and the accounts underneath.
Everyone is using AI a bit differently? Let's make it one way.
On a free call we ask what your team repeats every week, and tell you where these tools help and where they don't.