Skip to content
Compliance · Germany

DSGVO and Chatbots: the 2026 Compliance Checklist for German Businesses

German data protection officers look closely at chatbots. A practical 10-point checklist for running an AI chatbot under the GDPR (DSGVO), the BDSG and the TDDDG, with the questions to ask your vendor before launch.

Author
intoCHAT Team
Published
Last updated
Reading time
10 min read
Illustration: a data protection checklist, a shield with a padlock over a chat widget and a data processing agreement
On this page
  1. What the DSGVO requires for a chatbot
  2. Personal data in chatbot conversations
  3. Browser storage and the TDDDG
  4. Tell visitors they're talking to an AI
  5. Where data is stored and US transfers
  6. Auftragsverarbeitungsvertrag (AVV)
  7. Retention and data subject rights
  8. Access control and audit trail
  9. Data protection officer and impact assessment
  10. The practical 10-point checklist
  11. Sector-specific requirements
  12. How intoCHAT helps with the checklist
  13. Frequently asked questions

If your business operates in Germany and runs an AI chatbot, you are processing personal data under the EU GDPR (DSGVO in German) and the Federal Data Protection Act (Bundesdatenschutzgesetz, BDSG). German data protection authorities and company data protection officers (Datenschutzbeauftragte) look closely at chatbots, because they collect free text, often use AI providers outside the EU, and run on every page of a website. This article is a practical checklist for a website chatbot: what to decide, what to document and what to ask your vendor before launch. It is not legal advice.

What the DSGVO requires for a chatbot

Every processing activity needs a legal basis under Article 6. For a customer service chatbot, that is usually legitimate interest (Art. 6(1)(f)), backed by a documented balancing test, or Art. 6(1)(b) when visitors ask about an order, a contract or an offer. Marketing messages need consent (Art. 6(1)(a)). German DPOs expect the basis and the balancing test to be written down before launch, not copied from a vendor template.

You also have to inform visitors (Art. 13): who you are, why you process their data, who receives it, whether it leaves the EU and how long you keep it. The chatbot needs its own section in your Datenschutzerklärung.

Personal data in chatbot conversations

Almost every chat contains personal data. Visitors type names, email addresses, order numbers, and sometimes details about their health or finances. If data can be linked to a person, even indirectly, the GDPR applies. Health data and other special categories (Art. 9) need an extra legal basis, so a chatbot for a medical practice needs more care than one for a furniture shop.

The practical bar: collect only what you need, set a retention period, handle data subject requests (access, rectification, erasure, portability, objection), and describe the chatbot in your Datenschutzerklärung.

Browser storage and the TDDDG

Section 25 of the TDDDG (until May 2024 called TTDSG) governs storing information in a visitor's browser. Storage needs consent unless it is strictly necessary for a service the visitor explicitly asked for. A chat widget usually stores an ID so the conversation survives a page change. Ask your vendor exactly what its widget stores and when, and decide with your DPO whether it falls under the exception or belongs in your consent banner. Analytics or advertising cookies set by a chat tool always need consent.

Tell visitors they're talking to an AI

Since August 2026, Article 50 of the EU AI Act requires that people are told when they interact with an AI system, unless it is obvious. Introduce the agent as an AI assistant in its greeting, and link your Datenschutzerklärung from the chat window.

Where data is stored and US transfers

The GDPR does not require data to stay in Germany. Data may be processed in the EU and EEA, and in countries with an adequacy decision, such as Switzerland or the UK. Since July 2023, US companies certified under the EU-US Data Privacy Framework are covered by an adequacy decision too. Other transfers need standard contractual clauses and a transfer impact assessment, a requirement that goes back to the CJEU's Schrems II ruling.

For a chatbot, ask about two places:

  • Where conversations are stored at rest: the vendor's database and file storage.
  • Where messages are processed: the AI model provider that writes the replies. Most large providers (OpenAI, Anthropic, Google) process in the US. Ask which one your vendor uses, whether it is certified under the Data Privacy Framework or covered by standard contractual clauses, whether it trains on your data and how long it keeps it.

Auftragsverarbeitungsvertrag (AVV)

Article 28 GDPR requires a processor agreement between you and your chatbot vendor. In Germany it is called an Auftragsverarbeitungsvertrag (AVV). It should cover:

  • Categories of personal data and of data subjects.
  • Purposes and duration of processing.
  • The sub-processor list, and how you are told about new ones.
  • Technical and organisational measures (TOMs).
  • Breach notification: the vendor tells you without undue delay, so you can notify the authority within 72 hours where Article 33 requires it.
  • Support with data subject requests.
  • Audit rights.
  • Deletion or return of data at the end of the contract.

Ask for the AVV and the sub-processor list before you go live.

Retention and data subject rights

Set retention per data category and write it down. A common pattern:

  • Conversation transcripts: a few months, longer only for a documented reason such as abuse review.
  • Leads: as long as you need them for the enquiry or the customer relationship.
  • Anonymised statistics: no limit, because truly anonymised data is outside the GDPR.

Then check that your chatbot lets you act on it: delete a conversation, delete a lead, and export what you hold about a person. You answer a data subject request within one month (Art. 12(3)).

Access control and audit trail

Article 32 asks for appropriate security. For a chatbot that means: each team member gets only the access they need, two-factor sign-in is on, and changes to the agent, its sources and its integrations are recorded. Banks, insurers and larger companies usually also ask for single sign-on.

Data protection officer and impact assessment

Under section 38 BDSG, a company needs a data protection officer if at least 20 people regularly process personal data by automated means. If your chatbot processes special categories of data on a large scale, or makes decisions about people, check with your DPO whether a data protection impact assessment (Art. 35) is needed.

The practical 10-point checklist

  1. Signed AVV with sub-processor list, TOMs, breach notification and audit rights.
  2. Storage location and AI provider documented, with the legal basis for any transfer outside the EU.
  3. Legal basis per purpose: legitimate interest with balancing test or contract for service; consent for marketing.
  4. Datenschutzerklärung updated: the chatbot, categories of data, recipients, transfers, retention.
  5. Browser storage checked under section 25 TDDDG, and the consent banner adjusted if needed.
  6. AI disclosure: the agent introduces itself as an AI and the chat links the Datenschutzerklärung.
  7. Retention per data category, with a routine that deletes old conversations and leads.
  8. Data subject rights tested end to end: access, erasure and export.
  9. Team access limited by role, two-factor sign-in on, changes logged.
  10. Incident process agreed with the vendor, so you can meet the 72-hour deadline.

Sector-specific requirements

  • Banking and insurance (BaFin): outsourcing rules (MaRisk AT 9) and, since January 2025, the EU's Digital Operational Resilience Act (DORA) for ICT third-party providers. Expect a governance sign-off before launch.
  • Healthcare: health data is a special category (Art. 9), and doctors and other professionals bound by secrecy under section 203 StGB need a confidentiality commitment from the vendor. Keep clinical questions out of the chatbot unless your DPO has approved it.
  • Public sector: BSI IT-Grundschutz, procurement-specific contract terms and often stricter hosting requirements.

How intoCHAT helps with the checklist

No software makes a chatbot compliant on its own, and intoCHAT holds no certification such as SOC 2 or ISO 27001. What it gives you are the controls and the documentation for most of the points above.

  • AVV and providers. You can request a data processing agreement (AVV) at privacy@intochat.ai. The privacy policy lists the providers: Vercel (hosting and file storage), Supabase (database), Redis Cloud in the EU (rate limits), OpenAI, Anthropic and Google (AI replies), Firecrawl (reading your website), Resend (email) and Stripe (payments).
  • Hosting. The database that holds what intoCHAT stores is hosted in Switzerland, which has an EU adequacy decision. There is no choice of region, and some providers process data in the US and other countries, under standard contractual clauses where the law requires them.
  • Model and provider choice. Each agent writes its replies with Claude Haiku 5.5 from Anthropic by default, and GPT-6 Luna from OpenAI answers if Anthropic doesn't. Instead, you can pick GPT-6 Luna on every plan, a Gemini model from Google from Starter or a larger Claude model on Pro. OpenAI is still used to index your knowledge and for background features such as summaries, so list it in your Datenschutzerklärung whichever model you pick. intoCHAT doesn't train AI models on your data or your visitors' data.
  • Answers from your content only. You decide which pages, files and Q&A the agent learns from. Built-in rules make it say what it can't confirm instead of inventing an answer, and it never asks for passwords or card details in the chat.
  • Browser storage. The widget stores an ID in the visitor's browser to continue the conversation and, when you turn those features on, a list of earlier chats and whether the launcher is shown. It sets no analytics or advertising cookies, which makes the TDDDG assessment easier.
  • Deletion and export. Delete a conversation with its messages and files, delete leads and contacts, and export conversations, leads and contacts as CSV to answer an access request. With Let visitors delete their chat, visitors can delete their own conversation from the widget. There is no automatic deletion schedule, so plan a regular clean-up to match your retention rule.
  • Disclosure. Name the agent and write its greeting yourself, and use the widget's footer text to link your Datenschutzerklärung.
  • Identity verification. On every plan, your server signs each signed-in user's id, so the agent knows who it is talking to and nobody can pretend to be another customer. You can let only verified visitors chat.
  • Team controls. Roles (Admin, Editor, Viewer) and access to selected agents from Starter, two-factor sign-in for everyone, and an audit log of account changes on Pro (90 days) and Enterprise (365 days). The log records who changed agents, sources, integrations, API keys and team members and who downloaded exports, never message content. Single sign-on with SAML and custom roles are on Enterprise.
  • Where the agent answers. Limit it to your own domains and block countries you don't serve.

Our chatbot GDPR guide walks through legal basis, privacy notice, processors, cookies and retention step by step. For an overview, see the AI chatbot for Germany page.

This article is informational and not legal advice. For a binding interpretation of the GDPR, BDSG and TDDDG for your business, consult a German data protection lawyer or your Datenschutzbeauftragter.

Frequently asked questions

Does the GDPR apply to website chatbots?

In almost every case, yes. Chat conversations involve personal data such as free text that may include names, contact details or health and financial information, and the GDPR applies whenever data can be linked to a person, even indirectly. In Germany the Federal Data Protection Act (BDSG) and, for browser storage, the TDDDG apply as well. This is general information, not legal advice.

What is the legal basis for a chatbot under the GDPR?

For customer service questions, the usual basis is legitimate interest under Article 6(1)(f), backed by a documented balancing test, or Article 6(1)(b) when the chat serves a contract or a pre-contract request. Marketing messages need consent under Article 6(1)(a). Document the basis for each purpose before launch instead of relying on a generic vendor template.

Do I need a data processing agreement with my chatbot provider?

Yes. Article 28 GDPR requires a processor agreement, called an Auftragsverarbeitungsvertrag (AVV) in Germany. It covers the categories of data, purposes, duration, sub-processors, technical and organisational measures, breach notification, data subject rights, audit rights and deletion at the end of the contract. intoCHAT provides a data processing agreement on request at privacy@intochat.ai.

Can a chatbot use AI models hosted in the US under the GDPR?

Yes, with a valid transfer mechanism. Since July 2023, transfers to US companies certified under the EU-US Data Privacy Framework rely on an adequacy decision; for others, standard contractual clauses and a transfer assessment apply. Check whether the model provider trains on your data and how long it keeps it, and list it in your Datenschutzerklärung. For banking, insurance or healthcare, many DPOs set stricter rules.

Does intoCHAT make a chatbot GDPR compliant?

No software makes a chatbot compliant on its own, and intoCHAT holds no certification such as SOC 2 or ISO 27001. It gives you the tools for much of the checklist: an AVV on request, a database hosted in Switzerland (which has an EU adequacy decision), a choice of AI provider per agent (OpenAI, Anthropic or Google), deletion and CSV export of conversations and leads, a delete button for visitors, identity verification, team roles, two-factor sign-in, and an audit log on Pro. OpenAI indexes your knowledge whichever model you choose, so it belongs in your Datenschutzerklärung.

Back to Blog
  • Illustration: a padlocked shield, a vendor checklist and data flowing from a chat widget into a safe
    Compliance · Switzerland

    Swiss FADP and Chatbots: the 2026 Compliance Checklist

    The revised Swiss Federal Act on Data Protection (revFADP, revDSG) has applied since September 2023. A practical checklist for running an AI chatbot under it, with the questions to ask your vendor before you sign.

    9 min read

  • Illustration of five chat assistants facing a smartphone with a customer message, a clipboard with a bar chart beside them and the Alps behind
    Study · Switzerland

    Do AI Chatbots Understand Swiss German? We Tested 5 Models

    50 customer messages from 7 regions, 5 current AI models, 1,000 answers: every model understood written Swiss German almost as well as standard German. The differences are in how they reply.

    6 min read

  • Editorial illustration of a closed shop inside a browser window, with a warm chat light connecting it to a visitor on a laptop at night
    Lead generation

    The 11:47 PM Customer: Your Best Lead Is Still Awake

    Someone is on your pricing page tonight, one unanswered question away from leaving. Here is how to build an AI lead generation chatbot that helps them take the next step—with a reusable prompt, ten tests and a seven-day launch plan.

    10 min read