Can My Employer See My Claude Chats? On a Work Account, Yes

Yes, if you use Claude through your company's Team or Enterprise plan. Anthropic's help center says the organization's Primary Owner "manages your Work account and all associated data", and that includes a data export with your conversations, uploaded files and usage. Enterprise companies can also turn on the Compliance API, which feeds chats and files into their security tools. On your own Claude Free, Pro or Max account, your employer has no access to your chats at all. The work laptop you use is a separate question: monitoring software on it can record what you type into any website.

No admin, owner or manager can open your chats inside the Claude app. The access is an export or an API feed, taken on purpose by the one or two people who hold the top role.

This is the Claude part of our guide to whether your employer can see your ChatGPT history.

Who can read your Claude chats, by plan

Sources are Anthropic's own help pages, read on 27 September: "Who owns and manages the data of my team?", "Export your organization's data", "Roles and permissions", "Access the Compliance API", "Access audit logs" and "Use incognito chats" and "Claim and migrate accounts on your domain".

Claude Team: the Primary Owner can export every chat

Each Team or Enterprise organization has exactly one Primary Owner. In Anthropic's roles table, "Request data exports" is ticked for the Primary Owner only. Owners, Admins and Users cannot request one.

The export works like this:

• The Primary Owner opens Organization settings > Data and privacy and clicks Export Data.
• Anthropic emails a download link to the Primary Owner. It expires after 24 hours.
• The file holds the organization's conversations, files and projects.

Anthropic says members of a Team or Enterprise plan cannot export their own data. Only the Primary Owner can.

Owners on a Team plan also see usage analytics: chats per day, the share of members with at least one chat, and "top members by chats". These are counts. They show that you use Claude and how much, not what you asked.

Claude Enterprise: exports, the Compliance API and audit logs

Enterprise adds three things on top of the export.

The Compliance API. Anthropic describes it as a way to "programmatically pull activity feed events, chat data, and file content" across Claude, Claude Desktop, Claude Mobile, Cowork and Claude Code. Only the Primary Owner can turn it on, in Organization settings > API. Once it is on, security and compliance products can read your chats continuously, without anyone clicking Export.

Audit logs. Owners and the Primary Owner can export 180 days of audit events. The log records who did what, such as creating a chat or a project, with an IP address and device. Anthropic says the titles and content of chats are not in the audit log, only their IDs.

Retention settings. Owners can set how long chats are kept, with a minimum of 30 days. By default Enterprise keeps chats indefinitely.

Enterprise Admins and Owners also see usage analytics, and on usage-based plans a spend report with one row per person, per model, listing request counts and tokens. Again, those are numbers and not the text of your prompts.

Incognito chats and deleted chats

Incognito mode keeps a chat out of your history and out of Claude's memory. It does not hide it from your employer. Anthropic's incognito article says that on Team and Enterprise plans:

• • incognito chats are included in the organization data exports;
• • they are kept for 30 days for safety, or longer under the company's retention policy;
• • they are included in the Compliance API on Enterprise.

Deleting a chat helps with exports. Anthropic says messages, files and projects deleted by a user or by the retention setting are not included in exports started after the deletion. But a company that streams chats through the Compliance API may already hold its own copy from before you deleted it.

Personal Claude on a work computer

Your personal Claude account belongs to you. Your employer cannot request an export of it, and Anthropic's consumer terms do not give employers any role. On a Team or Enterprise account the situation is reversed: Anthropic acts as a data processor for your company, and your company's agreement decides the rules.

The device is the weak point. If you open claude.ai on a company laptop, the company can still see:

• the site you visit, from its web filter or network logs (the full page text needs a TLS inspection proxy, common in larger companies);

• what is on your screen or what you type, if the laptop runs monitoring software.

We checked the first point on a Mac with macOS 26.5.2 on 6 October 2026. The first message a browser sends to open a secure connection, the TLS ClientHello, carries the site name in plain text. This Python script builds that message for claude.ai the way a browser does:

import ssl
ctx = ssl.create_default_context()
incoming, outgoing = ssl.MemoryBIO(), ssl.MemoryBIO()
tls = ctx.wrap_bio(incoming, outgoing, server_hostname="claude.ai")
try:
    tls.do_handshake()
except ssl.SSLWantReadError:
    pass
hello = outgoing.read()  # the first bytes the laptop sends
print(len(hello), "bytes; site name readable:", b"claude.ai" in hello)

1514 bytes; site name readable: True

So a company network can log that you opened claude.ai and when, with nothing installed on the laptop. Your prompts travel inside the encrypted part. To see whether your company decrypts that part, read who issued the certificate your browser gets:

echo | openssl s_client -connect claude.ai:443 -servername claude.ai 2>/dev/null | openssl x509 -noout -issuer

issuer=C=US, O=Let's Encrypt, CN=YE2

That is a public certificate authority, so our network was not decrypting the connection. If the issuer names your company or a security vendor, a TLS inspection proxy can read the page text too.

The details for each of those are in can my employer see my browsing if I use a VPN, can my employer see my screen and how employers monitor remote workers.

A personal Claude account on your work email

This is the case most people miss. If you signed up for Claude Free, Pro or Max with your work address, your company can find that account once it buys Claude Enterprise.

Anthropic calls it domain claiming. After the company verifies its email domain, an Enterprise admin sees every personal account on that domain with its email address, plan type, creation date and last active date, and can export that list. The admin cannot read the chats at that stage. The admin can then claim the accounts, and you get an email with a deadline of at least 30 days and two choices:

• Merge and join. Your old conversations, projects and files move into the company's Enterprise organization, where the Primary Owner's export and the Compliance API cover them.

• Join fresh. You start a clean work account and nothing moves. Export your personal data from Settings > Account before the deadline if you want a copy.

If you do nothing, you get a fresh account and lose access to the personal one. A Pro or Max subscription is canceled with a prorated refund. Team plans can block new personal accounts on the company domain but cannot claim existing ones. The safe setup is a personal Claude account on a personal email address.

How to check what your company can see

Look at the account name. In Claude, open Settings > Account. A work organization shows its name and your role. If you are a plain member, someone else holds the Primary Owner role.

Check your email. A Claude account created with a work email is the one your company can claim. Use a personal email for a personal account.

Watch for a domain claim email. It means your company found a personal account on your work address.

Read your company's AI policy. Companies that turn on the Compliance API usually say so in their acceptable-use or data-loss policy.

Ask IT which plan you are on. Team has exports only. Enterprise may have exports, audit logs and a live feed.

Claude vs ChatGPT vs Copilot at work

The ChatGPT side is in can my employer see my ChatGPT history, and the Copilot side in can my employer see my Copilot chats. Chat tools work the same way: see can Slack admins see your DMs and can my employer read my Teams chats.

FAQ

Can a Claude Team admin see my chat history? Not in the app. An Admin has no export rights. The Primary Owner can export all chats, including yours.

Can Claude Enterprise see my chat history? Your company can, through the Primary Owner's export and, if it is turned on, the Compliance API.

Does Anthropic share my chats with my employer? Anthropic does not send your personal account's chats to anyone. On a work account the chats already belong to your company's organization, and Anthropic provides the export tools.

Can my employer see my Claude chats if I use my own phone? Yes, if you sign in to the work account. The chats live in the company organization, not on the phone.