What the BGH Schufa rulings mean for companies using scores

On 21 October the Federal Court of Justice (BGH) ruled on Schufa scoring and changed how much transparency people can demand about automated scores. Companies that buy scoring data or run automated assessments should read the rulings as guidance for their own processes.

Sandra Leistner4 min read

Illustration: A glass display case reveals the inner workings of a small scale, with a key and a wedge lying in front.

Why the rulings also affect users of scores

Most commentary treated the decisions as a dispute about Schufa, the German scoring agency, and about an old scoring model. A commentary on netzpolitik.org by the lawyer Beata Hubrig argues that this view is too narrow. In her reading the court described how individuals and algorithmic rating systems relate to each other under the GDPR in general.

The commentary notes that Schufa has been moving gradually to a new model since spring, which is why many readers saw the rulings as outdated. The author rejects that view, because the rulings set the standard for any later automated assessment.

That matters for the companies on the other side of the score: banks, landlords, telecoms providers, online shops and insurers. They receive a value, act on it and may later face a request from the person concerned. The commentary also points out that a new scoring provision in the Federal Data Protection Act (BDSG) takes effect on 20 November and follows the GDPR closely. Our page on GDPR and AI covers the wider data protection frame.

What the BGH decided

The commentary describes two parts. First, the court eased the burden of proof. If a score is delivered and a contract is refused shortly afterwards, there is a factual presumption that the score triggered the refusal. The person concerned no longer has to produce internal documents of the lender.

Second, the court rejected a general right to learn how Schufa calculates and weights its criteria, one that would need no particular reason. A deeper right to information arises when a score has had noticeable consequences for a contract. The lower court in Dresden had gone further, and the BGH overturned that line.

The background is a December 2023 ruling of the European Court of Justice. It held that creating a score can itself be an automated decision under Art. 22 GDPR if the score plays a decisive role in whether a contract is concluded. The BGH rulings build on that and decide how it is proven and how far the right to information reaches.

What this means for your decision processes

The presumption works against the user of a score. If your refusal follows a score closely in time, you should expect a court to treat the score as decisive. Whether Art. 22 GDPR then applies to your process depends on how much room your staff really have to deviate from the score.

The right to deeper information arises once a concrete disadvantage has occurred. The commentary also names other uses in view: automated insurance scoring, fraud detection in online retail and screening of rental applicants.

The team that owns the process is affected first, usually the one handling lending, sales or property management. Data protection officers and, where one exists, the works council (Betriebsrat) are affected as well, because automated assessments of staff follow the same logic.

Preparing access requests in five steps

  1. List every process in which a score, a risk value or an automated rating feeds into a contract decision, including the supplier and the date of each query.
  2. Record for each process whether a person reviews the result and whether that person can and does deviate from it.
  3. Keep the score, the decision and the date together, so the time link between them can be shown or explained.
  4. Ask your scoring providers which information about criteria and weighting they supply to you, and file their answers.
  5. Define who answers a request, within which time limit and from which documents, and test the route with a mock request.

None of this replaces legal advice on your own case. It gives your counsel a clean basis if a request or a dispute arrives, and it makes the internal routes clear before the first request comes in.

Answering requests from company documents with evidence

Answering such a request means finding contracts, e-mails, process descriptions and decision notes. These are often spread over several systems. souveraen.ai is an AI platform for company knowledge. It answers questions from connected sources and cites the document, version and page for each answer, so your team can check what a reply rests on. More on this is on the page about answers with sources.

It connects SharePoint, OneDrive, Outlook and file servers read-only and with the existing permissions. It runs either On Demand, hosted in Germany, or on-premise, with the same software in both modes. Further background is on the security page.

Sources

More articles

All articles

How safe are AI API routes via resellers and routers?

Security firm Dragos sent about two million requests with fake credentials to AI providers and found compromised services and resellers reading customer prompts. Companies that use AI through third parties need to know who sits on the path of their data.

Data protection4 min

Run Text Generation Inference with Docker on Your Server

Text Generation Inference (TGI) serves open language models from a single Docker container on your own hardware. This tutorial starts it, sends requests over HTTP and keeps prompts inside your network.

Tutorial4 min

Cloud outage liability: who answers when Microsoft fails

On 12 November 2026 a software fault took Exchange Online and Outlook offline for about 36 hours. A legal analysis published shortly afterwards argues that responsibility for such outages often lies with the operator rather than the provider.

Regulation4 min

Test it with your own documents.

Free for 3 days, then from €24.90 per user per month. Cancellable monthly.

Free for 3 days, then from €24.90 per user per month.Free for 3 days
Try free for 3 days