To use ChatGPT in a GDPR-compliant way, you need more than one privacy switch. Tuesday morning: a colleague in sales pastes a customer’s email and order history into a personal ChatGPT account, and nobody knows whether that was allowed. Whether ChatGPT can be compliant at all is answered in Is ChatGPT GDPR-compliant?; this guide shows how to set it up, step by step, as of November 2026, and how Superkind’s ready-made AI team for work with customer data takes over the tasks with personal data. It is not legal advice; confirm the details with your data protection officer.
Key takeaways
Get a ready-made AI team - Superkind brings an AI team for work with customer data (Inbox Assistant, Lead Sorter, Onboarding Helper, Call Note Taker) running automatically in Outlook, Teams and Slack. Start for free with unlimited users, or build your own.
Keep control, keep the data in Europe - hosted and processed in Europe, GDPR compliant, no training on your data, and you decide what needs approval.
ChatGPT: a business plan, not a personal one - Business, Enterprise and the API exclude your data from training and come with OpenAI’s DPA; personal plans do not12.
EU data residency is Enterprise and API only - for new Enterprise and Edu workspaces and API projects, not for Business4.
Rules beat settings - a written team rule plus AI literacy training is required under Art. 4 of the EU AI Act since February 202510.
Your AI team for work with customer data
What does it mean to use ChatGPT in a GDPR-compliant way?
Using ChatGPT in a GDPR-compliant way means running it on a plan with a signed data processing agreement, training off, a known place of processing, set retention and a written rule on what staff may put in. The GDPR does not ban the tool; it makes your company, as controller, responsible for the personal data that goes in9. One example of the strongest setup is Superkind: an AI team that does tasks with customer data in Outlook, Teams and Slack, hosted in Europe.
How it works with Superkind: your AI team for work with customer data
You write to Superkind by email, in Teams, in Slack or in the web app, like to a colleague, and Superkind does the task in your programs: 1,700+ integrations, including HubSpot, Zendesk and Personio. It works next to ChatGPT instead of replacing it, hosted in Europe and GDPR compliant.
With Superkind you don’t start from zero: it comes with ready-made AI employees for work with customer data. They run automatically in the background, and you decide where they ask you first. Use them right away, adapt them, or build your own.
Inbox Assistant
AI employee for customer emails
Customer emails carry names, addresses and order numbers, exactly what should not end up in a personal chat account. When a new email reaches support@, the Inbox Assistant finds the order in Shopify, checks earlier messages and drafts the reply in your tone. Refunds over EUR 200 wait for your OK, everything else goes out on its own.
Lead Sorter
AI employee for lead qualification
Web form leads come in overnight with names, phone numbers and budgets. Every morning the Lead Sorter enters them in HubSpot, checks them against your criteria and posts the ones worth a call in Slack. Nobody copies contact data into a chat window.
Onboarding Helper
AI employee for employee onboarding
Employee data is personal data, and onboarding touches a lot of it. As soon as a new hire is in Personio, the Onboarding Helper builds the checklist, asks IT for the laptop, schedules the welcome email and posts the plan in Teams.
Call Note Taker
AI employee for call notes
After a call, the notes sit in someone’s head or in a private chat. The Call Note Taker writes them into the HubSpot deal after every sales call and suggests the next step, so the customer history stays in your CRM and not in a chat tool.
After every sales call: write the notes into the HubSpot deal and suggest the next step.
Reasoning ⌄Call with Brandt Logistik logged in HubSpot: 3 shops, budget approved, decision by 5 December. Follow-up email drafted for Monday.
Runs automatically after every call
Every employee is ready to use right away. Or build your own on the platform.
Safe by default
Customer emails, leads and employee data stay in Europe and never train a model, and the DPA comes through the Trust Center.
- Hosted and processed in Europe
- GDPR compliant
- Aligned with the EU AI Act
- No training on your data
- DPA through the Trust Center
What you need
Checklist before you start
- One person who owns AI in the team and may sign the DPA
- A list of what your team uses ChatGPT for, marked where personal data is involved
- Your data protection officer for the lawful basis and the record of processing
- Admin access to your ChatGPT workspace, or a budget for ChatGPT Business or Enterprise
- For tasks with customer data: a free Superkind account at app.superkind.ai, connected to your programs via 1,700+ integrations such as Pipedrive, Freshdesk and Google Drive
Step by step: ChatGPT GDPR-compliant in six steps
Step 1: Pick the right plan
Move every work use to ChatGPT Business, Enterprise or the API. On personal plans, chats can train OpenAI’s models unless each user opts out, and there is no business DPA15.
| Plan | Training on your data | DPA | Data in Europe |
|---|---|---|---|
| Superkind | Never | Yes, via the Trust Center | Yes, hosted and processed in Europe |
| ChatGPT Free, Plus, Pro | On by default, opt-out | No business DPA | No |
| ChatGPT Business | Off by default | Yes, via OpenAI’s DPA form | Not listed by OpenAI |
| ChatGPT Enterprise | Off by default | Yes | New workspaces, on request |
| OpenAI API | Off by default | Yes | New projects, on request |
As of November 2026 (Superkind: October 2026). ChatGPT Business costs USD 25 per user per month, or USD 20 with annual billing, for at least two seats7.
Step 2: Sign OpenAI’s DPA
Art. 28 GDPR requires a processing contract with every provider that handles personal data for you9. OpenAI signs a DPA for Business, Enterprise and the API: complete the DPA form linked on its DPA page23. Companies in the EEA or Switzerland contract with OpenAI Ireland Ltd. File the signed DPA with your record of processing.
Step 3: Check where your data is processed
OpenAI offers European data residency (EEA and Switzerland) for new Enterprise and Edu workspaces and new API projects; inference residency keeps the processing in Europe too14. Existing workspaces cannot be moved, and workspace metadata, billing and logins stay outside the region4. For Business, OpenAI lists no data residency, so plan with processing outside Europe under the DPA’s transfer terms. Owners see the region under Workspace settings, General.
Step 4: Switch off training, set retention and apps
- Personal plans - Settings, Data controls, switch off “Improve the model for everyone”. Temporary Chats are not used for training and are deleted after 30 days5.
- Business and Enterprise - excluded from training by default1.
- Retention - deleted chats are removed within 30 days unless the law requires otherwise; on Enterprise, admins set how long chats are kept26.
- Apps - on Business, connected apps are on by default. Switch off in the admin console what the team does not need8.
Step 5: Write the team rule and train the team
A setting does not stop anyone pasting a customer file. Write one page on what never goes into ChatGPT:
- Personal data - customer and staff names, emails, order and HR data, unless plan, DPA and lawful basis cover it.
- Sensitive data - health, religion or union membership, never.
- Secrets - passwords, API keys, contract terms, unpublished figures.
- Unchecked output - nothing reaches a customer before a person has read it.
Since 2 February 2025, Art. 4 of the EU AI Act requires AI literacy among staff who use AI10; a short training on this rule is the start. Our guide to writing a company AI policy has a template, and ChatGPT at work covers what is allowed.
Step 6: Take the European path for sensitive work
Tasks that touch customer or staff data every day, such as support emails, leads or onboarding, are easiest with an AI employee processed in Europe from the start. Superkind uses the same model class (OpenAI, Anthropic, Mistral and Google), processed in Europe, with no training and the DPA through the Trust Center:
Answer Ms Berger’s complaint about order 4012 using the order in Shopify. Refunds over EUR 200 wait for my OK.
Superkind reads the order in Shopify, drafts the reply in Outlook and sends it after you click Approve. More in email inbox triage and in our guide to switching to European AI.
Set it up as a routine
Every new email to support@: draft the reply from the Shopify order and earlier tickets. Refunds over EUR 200 wait for my OK, everything else goes out.
From then on it runs as a routine, and every case adds to your Company Brain: customers, prices and how your team answers. When the process grows, you build your own app on the Superkind platform, and it belongs to you.
Data protection and EU hosting
- Approval where you want it - you decide which steps wait for your OK before Superkind sends, books or changes something.
- Sovereign and independent - hosted and processed in Europe under European law, no lock-in to one model vendor. Your knowledge, routines and apps belong to you.
- Any program - no integration? Superkind works natively via computer use, in the browser or desktop client, inside your own infrastructure.
Other ways
| Way | Effort | Data protection risk | Best for |
|---|---|---|---|
| Superkind | Low: write the task, connect programs | Low: Europe, DPA, approval | Daily tasks with customer or staff data |
| ChatGPT Enterprise | Medium: sales contract, setup | Low with EU residency and rule | Teams that want ChatGPT itself |
| ChatGPT Business | Low: self-serve, per seat | Medium: outside Europe, rule needed | Drafting and research without personal data |
| Personal accounts | None | High: training on, no DPA | Private use only |
| Ban on AI | None | High: staff use it anyway | Nobody |
As of November 2026. More options in European ChatGPT alternatives and AI tools from Europe.
Common mistakes
- Staying on personal accounts - a Plus account paid by an employee has no business DPA, even for work use.
- Assuming Business means EU hosting - OpenAI lists EU data residency only for Enterprise, Edu and the API1.
- Forgetting the old workspace - residency applies only to new workspaces and projects4.
- Leaving every app on - a connected mailbox brings in more personal data than planned.
- A ban instead of a rule - bans push staff to private tools (see the shadow AI playbook).
6 tips for using ChatGPT in a GDPR-compliant way
1. Sort tasks before you pick tools
Split them into “no personal data” (ChatGPT Business is fine) and “personal data” (the European path).
2. Replace names before you paste
“Customer A, order X” drafts a reply as well as the real name.
3. Give the setup one owner
Plan, DPA, apps and team rule belong in one hand.
4. Re-check terms once a year
OpenAI updates DPA and plans; the current DPA dates from 1 January 20263.
5. Put approvals where the risk is
Refunds, contracts and HR messages wait for a person; routine replies run on their own.
6. Make data protection part of AI literacy
Thirty minutes per quarter on the DPA, where data is processed and who can see it covers Art. 4 of the EU AI Act.
Conclusion
ChatGPT is GDPR-compliant in use when plan, DPA, place of processing, retention and team rule fit together. For drafting without personal data, ChatGPT Business or Enterprise does the job; for daily work with customer and staff data, Superkind is the shorter path: processed in Europe from the start and working in your programs through 1,700+ integrations such as Shopify, HubSpot and Personio. Next step: list the three tasks with the most customer data and give the first to Superkind today (pricing).
Frequently Asked Questions
Move all work use to ChatGPT Business, ChatGPT Enterprise or the OpenAI API, sign OpenAI’s data processing agreement, check where the data is processed, set retention and connected apps, and write a team rule on what never goes in, backed by AI literacy training under Art. 4 of the EU AI Act. For tasks with customer or employee data, the simplest path is an AI employee like Superkind that is hosted and processed in Europe from the start. This is not legal advice; confirm the details with your data protection officer.
Not for personal data at work. On the personal plans (Free, Plus, Pro) chats can be used to train OpenAI’s models unless each user switches “Improve the model for everyone” off, and there is no business data processing agreement. For texts without personal data, with training switched off, the risk is low, but customer and staff data belong on a business plan or a European tool. As of November 2026.
OpenAI offers European data residency (EEA and Switzerland) for new ChatGPT Enterprise and Edu workspaces and for new API projects, with optional in-region inference on top. OpenAI does not list data residency for ChatGPT Business, so Business data is processed mainly outside Europe under the transfer terms of the DPA. Check the current terms with OpenAI before you rely on it. As of November 2026.
Your team writes to Superkind in Outlook, Gmail, Microsoft Teams, Slack or the web app, and Superkind does the task in your programs through 1,700+ integrations, for example Gmail, Notion and Aircall. Ready-made AI employees such as the Inbox Assistant, Lead Sorter, Onboarding Helper and Call Note Taker run automatically, and you decide which steps wait for your approval. It uses models from OpenAI, Anthropic, Mistral and Google, processed in Europe, and keeps what it learns in your Company Brain.
Yes. Superkind is hosted and processed in Europe, GDPR compliant, aligned with the EU AI Act (Regulation (EU) 2024/1689) and never trains models on your data. The data processing agreement comes through the Trust Center at trust.superkind.ai. You still name the lawful basis for your own processing, as with any provider.
ChatGPT Business costs USD 25 per user per month, or USD 20 with annual billing, for at least two seats (as of November 2026). Superkind starts for free with a one-off credit of up to EUR 150, no credit card and unlimited users; the Business plan starts at EUR 50 per month for the whole company, never a price per seat. EUR 100 covers up to 1,000 tasks or 5,000 questions. Superkind prices as of October 2026.
Related Articles
- Is ChatGPT GDPR-Compliant? Examples, Costs and How to Start
- European ChatGPT alternatives
- How to switch to European AI
- AI tools from Europe
- ChatGPT at Work: What’s Allowed, Forbidden, and Trained
- Shadow AI: The Governance Playbook
Sources
- OpenAI - Business data privacy, security, and compliance (no training by default, data residency eligibility)
- OpenAI - Enterprise privacy (DPA for Business, Enterprise and API, retention controls)
- OpenAI - Data Processing Addendum (v. 1 January 2026)
- OpenAI Help Center - Data residency and inference residency for ChatGPT
- OpenAI Help Center - Data controls in ChatGPT (training opt-out, Temporary Chat)
- OpenAI Help Center - Chat and file retention in ChatGPT
- OpenAI Help Center - Managing billing and seats in ChatGPT Business
- OpenAI Help Center - Admin controls, security, and compliance for plugins and apps
- GDPR - Article 28: Processor
- EUR-Lex - Regulation (EU) 2024/1689 (EU AI Act), Article 4: AI literacy
Send Superkind your first customer email.
Start for free in Outlook, Teams or Slack, processed in Europe. Or let us show you in 30 minutes how it runs in your systems.

