Claude

Does Claude train on your company data? Privacy on Team, Enterprise and personal plans

Does Anthropic use your conversations to train Claude? How privacy differs between Claude Team, Enterprise, the API and personal plans, and the rules to give staff.

It is the first question every security team asks before approving Claude: will Anthropic use our data to train its models? The answer depends on which plan your people are using. For business plans the default is clear. For personal accounts, it depends on settings each user controls.

This guide explains the difference in plain language and gives you the rules to put in front of staff. It is general information, not legal advice. Anthropic’s privacy terms are the source of truth and can change, so review them before you decide.

Business plans: Team, Enterprise and the API

Anthropic’s commercial terms cover Claude Team, Claude Enterprise and the Claude API (including access through cloud providers). Under those terms, Anthropic does not train its models on your inputs and outputs by default. Your company is the customer, and the data belongs to your organisation.

Business plans also give your admins control: who has access, which connectors are enabled and, on Enterprise, retention settings and audit logs. That control is as important as the training question, because it decides what data Claude can reach in the first place.

Personal plans: Free, Pro and Max

Personal accounts are governed by Anthropic’s consumer terms. Since late 2025, consumer users choose in their privacy settings whether their chats can be used to improve Anthropic’s models, and the choice affects how long data is kept. That is reasonable for individuals, but it means your company has no control over data an employee puts into a personal account.

The practical rule: company data goes into a company-managed business plan. Personal accounts are for personal use.

What “not used for training” does and doesn’t mean

  • It means your conversations don’t become part of a future model that other customers use.
  • It does not mean data is never stored. Conversations are kept so users can see their history, under the retention rules of your plan.
  • It does not replace your own controls. Confidential client data, personal data and secrets still need rules about where they may go.
  • It does not cover third-party tools. If someone pastes Claude output into another app, that app’s terms apply.

Comparison at a glance

Personal (Free, Pro, Max)Business (Team, Enterprise, API)
TermsConsumerCommercial
Used to train modelsUser’s choice in settingsNo, by default
Who controls accessThe individualYour admins
Connectors to company systemsPersonalManaged by admins
Fit for client or company dataNoYes, under your policy

Five rules to give your staff

  1. Use your company Claude account for any work task.
  2. Never paste passwords, API keys or payment card data, in any tool.
  3. Client personal data only where your policy allows it, and only in the company account.
  4. Check facts, numbers and sources before you send or publish anything Claude wrote.
  5. If unsure, ask the named contact in the AI policy before you paste.

Our AI usage policy template puts these rules on one page.

How to roll this out safely

Most data incidents with AI tools come from people, not vendors: the wrong file in the wrong tool. The fix is a clear policy, an approved business account and training with real examples. In our Claude for Teams sessions, the safe-use module uses your own documents to show what can go where, and our AI literacy session gives every employee the same baseline.

Frequently asked questions

Does Claude train on Team account data?

By default, no. Anthropic does not use inputs and outputs from its commercial products, including Claude Team, Claude Enterprise and the API, to train its models.

Is the free version of Claude safe for company data?

Personal plans (Free, Pro and Max) are governed by consumer terms and their own privacy settings, which the user controls. For company data, use a business plan your company manages.

Can we stop employees using personal Claude accounts?

You can set a policy and provide an approved business account. Blocking alone tends to push use onto personal devices, so give people a good approved tool and train them on it.

Vishal Wadhwani

Vishal Wadhwani
Founder of Automate Everything. Trains teams on Claude, ChatGPT, Copilot and n8n, and builds the automations that follow. LinkedIn

Next step

What would your team stop doing by hand if it knew how?

Tell us which teams and which tools. Within 48 hours you get a programme, dates and a fixed price. The call is 30 minutes, with the trainer who would run your sessions.

4.7 Excellent · 31 Trustpilot reviewsVishal Wadhwani on LinkedIn