English · Türkçe · Português · Español · Deutsch · Bahasa Indonesia · Filipino
© 2026 FirstPrompt · the purchase counts on this page are live
Privacy
Your team works in your company's own Claude. FirstPrompt keeps what the training and your dashboard need. Here is each item.
Questions your IT team will askLast reviewed October 8, 2026
Legal review pending. Until a lawyer has read it, this page describes how the product works. It is not legal advice.
Your team uses your company's own Claude accounts. FirstPrompt never connects to them and cannot see what happens in them.
The workflows
The prompts people paste and the schedules they set run in Claude, not on FirstPrompt.
Files, email and connected tools
What Claude reads to do a task goes to Anthropic, under your company's Claude terms.
Drafts and results
What Claude writes lands in your team's Claude and email: drafts, tables, summaries.
Chats and sessions
Kept by Anthropic under your plan's terms. On Team and Enterprise, Anthropic does not train on them by default.
Your company owner never sees exercise answers or tutor chats. Two items also go to Anthropic, as marked.
Accounts
Each person's name and work email, their department and their role.
Lesson progress
Which lessons each person has finished.
Readiness answers
Department, AI use, tools, what to keep out of AI, and up to 3 task names (120 characters each) with minutes per week and the workflow each one matched. Task names people type are included.
Typed task names
Outside the Turkish check, a visitor's typed tasks are matched to a workflow by Claude, through FirstPrompt's API account. Otherwise they are matched by keywords and not sent.
Check-in minutes
Once a month, how long each task takes now.
Workflow statuses
Set up or running, as each person marks it. We never see a run.
Exercise answers
What people write in lesson exercises.
Tutor chats
Conversations with Pip, the lesson tutor. Pip is Claude, made by Anthropic.
FirstPrompt
From firstprompt.build/companies/privacy · Last reviewed October 8, 2026
Legal review pending. Until a lawyer has read it, this page describes how the product works. It is not legal advice.
For your IT team
Short answers for whoever approves new tools at your company. Print this part, or save it as a PDF, and send it on.
Your team uses your company's own Claude accounts. FirstPrompt never connects to them and cannot see what happens in them.
The prompts people paste and the schedules they set run in Claude, not on FirstPrompt.
Cowork works on your files by sending what Claude reads to Anthropic. Connect only the folders a task needs.
What FirstPrompt stores: each person's name and work email, lesson progress, up to 3 task names (120 characters each) with minutes per week, exercise answers, and tutor chats.
Also each person's monthly check-in minutes and which workflows they set up or run. We never see a run.
Your company owner never sees exercise answers or tutor chats.
The lesson tutor is Claude, made by Anthropic. What you type to it is sent to Anthropic through FirstPrompt's own API account. Anthropic does not train on it and deletes it within 30 days, unless it is flagged for a usage-policy violation or the law requires otherwise.
The readiness check sends tasks a visitor types themselves to Claude the same way, through the same account, to match each one to a workflow. A suggestion someone picks is matched without sending anything.
Do not type personal data or customer data into exercises or the tutor.
At setup, your owner picks what never goes into AI: technical drawings, prices, customers' personal data, employee records or company financials.
Every workflow is rewritten without it before anyone starts. Where a task is about a drawing, the safe version refers to it by its number and never opens it. A workflow that cannot work without it is taken out of the plan.
Each prompt carries the rule as a line Claude follows, and the same rule is in the AI usage policy your team accepts when they join.
A rule is an instruction and a policy, not a technical lock. It cannot stop someone attaching a file by hand.
On Claude Team and Enterprise, Anthropic does not train its models on your content by default. The exception is feedback people choose to send, which your Claude owner can switch off.
Do not use personal Claude Free, Pro or Max accounts for company work. On those plans chats can be used for training unless the user turns it off.
Under Anthropic's Commercial Terms your company is the data controller, and Anthropic processes data for you under its Data Processing Addendum.
Anthropic stores data in the United States and may process it in the US, Europe, Asia and Australia. There is no Turkey-only or EU-only option.
Your company's chats and sessions in Claude are kept by Anthropic under your Claude plan's terms, not by FirstPrompt.
What people type to the tutor is deleted by Anthropic within 30 days, unless it is flagged for a usage-policy violation or the law requires otherwise.
FirstPrompt keeps tutor chats with the account. We have not set a deletion period yet. When we do, it will be stated here.
The tutor
The lesson tutor is Claude, made by Anthropic. What you type to it is sent to Anthropic through FirstPrompt's own API account. Anthropic does not train on it and deletes it within 30 days, unless it is flagged for a usage-policy violation or the law requires otherwise.
The readiness check sends tasks a visitor types themselves to Claude the same way, through the same account, to match each one to a workflow. A suggestion someone picks is matched without sending anything. Typed tasks are matched by keywords instead, and not sent, in the Turkish check, for members of a company, and once the check has used Claude a set number of times in the past hour. If Claude does not answer within a few seconds, the keyword match decides.
Other AI parts of the lessons work the same way: prompts typed into a lesson's practice terminal, and the text of a project page someone submits for review.
FirstPrompt keeps tutor chats with the account. We have not set a deletion period yet. When we do, it will be stated here.
Do not type personal data or customer data into exercises or the tutor.
Your company owner
Sees
Never sees
Your Claude plan
A company plan: Team or Enterprise
Do not use personal Claude Free, Pro or Max accounts for company work. On those plans chats can be used for training unless the user turns it off.
No training by default
On Claude Team and Enterprise, Anthropic does not train its models on your content by default. The exception is feedback people choose to send, which your Claude owner can switch off.
Your company is the controller
Under Anthropic's Commercial Terms your company is the data controller, and Anthropic processes data for you under its Data Processing Addendum.
Where Anthropic processes data
Anthropic stores data in the United States and may process it in the US, Europe, Asia and Australia. There is no Turkey-only or EU-only option.
Connect only what a task needs
Cowork works on your files by sending what Claude reads to Anthropic. Connect only the folders a task needs.
Questions
Questions about this page, or a request about your data:
support@firstprompt.build