Skip to content
Zrico
Menu

For every firm

Is Claude Safe for Client Data?

How Anthropic handles your data in Claude Free, Pro and Max versus Claude Team, Enterprise and the API, and what that means for client confidentiality.

Updated 7 min read

“Is Claude safe?” depends on which Claude you mean. Anthropic offers personal plans and business products with different data terms, and what matters for a law or accounting firm is which one you use, how it is set up, and what your professional rules require. This guide sets out what Anthropic’s own pages say, as of October 2026.

Two families of products

Anthropic separates its products into two groups.

Consumer products. These are the personal Claude plans: Free, Pro and Max. Each person signs up and accepts the terms individually.

Commercial products. These include Claude for Work, which covers the Team and Enterprise plans, and the Anthropic API. They are governed by Anthropic’s Commercial Terms.

The data rules differ between the two groups. The rest of this guide takes them in turn.

Personal plans: training depends on a setting

Anthropic’s privacy center says it uses chats and coding sessions from Free, Pro and Max to improve its models if you choose to allow it in the Model Improvement setting. It may also use conversations flagged for safety review, and those you share through programs you opt into.

If Model Improvement is on, Anthropic says it may keep new or resumed chats in de-identified form for up to 5 years in its training pipelines. If you turn it off, Anthropic says it will not use your previous or new chats for future training. Data already in training runs or trained models is not removed.

There is one exception to know about. If you give thumbs-up or thumbs-down feedback, Anthropic says the entire related conversation is stored for up to 5 years and may be used for training, with the link to your user ID removed.

Incognito chats

Anthropic says incognito chats are not used to improve Claude, even if Model Improvement is on.

Deleting chats

When you delete a chat, Anthropic says it disappears from your history right away and is deleted from its back-end systems within 30 days. If its automated systems flag a chat for a usage policy violation, Anthropic may keep the inputs and outputs for up to 2 years, and safety classification scores for up to 7 years. It may also keep data longer where the law requires.

Commercial products: no training by default

Anthropic’s Commercial Terms say Anthropic “may not train models on Customer Content from Services.” Its privacy center says that by default it will not use inputs or outputs from commercial products to train its models.

The exception is feedback. If you explicitly report feedback or bugs, for example with the thumbs-up or thumbs-down buttons, or otherwise choose to allow it, Anthropic says it may use those chats to train its models. Tell staff not to send feedback on client conversations.

The Commercial Terms also say:

  • The customer keeps all rights to its inputs and owns its outputs.
  • Data is processed under Anthropic’s Data Processing Addendum, which the terms incorporate.

For its commercial products, Anthropic lists SOC 2 Type I and Type II reports, ISO 27001:2022 and ISO/IEC 42001:2023 certifications, and a HIPAA-ready configuration with a business associate agreement available.

Claude Team

Claude Team is Anthropic’s self-serve business plan. Its pricing page says it is for teams of 2 to 150, at $20 per Standard seat per month billed annually or $25 monthly, and $100 or $125 for a Premium seat. It includes single sign-on and states “No model training on your content by default.”

Anthropic says chats in its business products are kept so people can return to them, and deleted chats are removed from back-end storage within 30 days. Custom retention settings are listed for Enterprise only. Anthropic’s help center says Team plans cannot enable HIPAA.

Claude Enterprise

Claude Enterprise is priced at $20 per seat per month billed annually, plus usage at API rates, with a 20-seat minimum. Anthropic lists SCIM, audit logs, a Compliance API and custom data retention controls. Owners can set retention periods, with a minimum of 30 days. Enterprise plans can enable a HIPAA-ready configuration and sign a business associate agreement.

The API

The API is for building your own applications. Anthropic says it deletes API inputs and outputs within 30 days of receipt or generation, unless a chat is flagged for a usage policy violation, the law requires longer, or you have agreed otherwise, such as under a zero data retention agreement.

API customers can also choose where inference runs. Anthropic’s documentation says the default lets inference run in any available geography, and a setting restricts it to US infrastructure for supported models.

What this means for client confidentiality

No training is an important protection, but it does not settle the question by itself. Your professional rules ask broader questions.

For lawyers. ABA Formal Opinion 512 says that before entering information relating to a representation into a generative AI tool, you must assess the risk that people outside the firm could access it. For tools that learn from inputs, it says you need the client’s informed consent first. A personal Claude plan with Model Improvement on fits that risk. With it off, it is still an individual’s account, with no firm control.

The terms matter in court too. In United States v. Heppner, a federal court held that a defendant’s own chats with Claude were not privileged, pointing in part to the Anthropic consumer privacy policy then in effect, which the court read as allowing training and disclosure to third parties. See our Heppner guide.

For accountants. The AICPA Code says a member should not disclose confidential client information without the client’s specific consent. Before sharing it with a third-party service provider, the member should have a confidentiality agreement with the provider, with reasonable assurance about its safeguards, or get the client’s specific consent. A business plan under the Commercial Terms can support the first route. A personal account generally cannot. Tax return information has its own federal rules; see our section 7216 guide.

For everyone. Even on a business plan, you should know:

  • how long chats are kept, and who sets that
  • where requests are processed
  • whether staff can send feedback that allows training
  • what your contract and Data Processing Addendum say

A sensible approach

  1. Do not use personal Claude accounts for client work. Say so in your AI policy.
  2. If you want to use Claude directly, use Team or Enterprise, which the firm controls.
  3. If you need retention controls, audit logs or a business associate agreement, check that your plan includes them; Anthropic lists them for Enterprise.
  4. Tell staff not to send feedback on client conversations.
  5. Re-read Anthropic’s privacy center at least twice a year. Its articles are dated, and they change.

Whatever the plan, Claude can be wrong. A professional must review its output before relying on it.

How Zrico fits

Zrico runs Claude through Anthropic’s commercial API, with inference restricted to the US. Anthropic keeps API data for up to 30 days and does not train on it; Zrico does not have a zero data retention agreement with Anthropic. Each firm gets its own private workspace, with its own application and databases, and Claude is the default. The firm owner can also turn on GPT and Gemini, whose processing is not limited to the US. The owner gets a usage and audit report and retention settings. Zrico has no SOC 2, ISO or HIPAA certification, does not make a firm compliant, and its output must be reviewed by a professional.

Sources

General information, not legal advice. Anthropic updates these pages often, so check the current versions and your jurisdiction’s rules.

Give your firm a private place to use AI.

$39 per seat per month after a 30-day free trial, 5-seat minimum. Your workspace is ready minutes after you sign up.

Your card is charged when the trial ends. Cancel any time before then and you pay nothing.

Start 30-day free trial