Request a consultation

focus area · obligations for online services

Digital Services Act: What Companies Have to Implement

Since February 2024 the Digital Services Act has regulated the obligations of online intermediary services EU-wide — from hosting providers to marketplaces. Not only large platforms are affected: anyone who hosts user-generated content, runs an online marketplace or allows reviews and comments has obligations. We help with classification and implementation — with AI where it reduces effort.

Who falls under the DSA

The DSA applies to intermediary services offered to users in the EU — regardless of where the provider is based. The obligations are tiered: mere conduit and caching have the fewest, hosting services more, online platforms even more; very large platforms and search engines face additional rules.

Two categories matter most for mid-sized companies: hosting services (third-party content is stored) and online platforms (third-party content is disseminated publicly — marketplaces, forums, review portals, community features). Micro and small enterprises are exempt from some platform obligations, but not from the basic ones.

The key obligations at a glance

Depending on classification, these include:

Points of contact

A central point of contact for authorities and for users, easy to find.

Terms and conditions

Understandable moderation rules, including information on automated procedures used.

Notice and action procedures

Users can report illegal content; reports are processed and answered traceably.

Statement of reasons

Anyone removing content or suspending accounts must give reasons to those affected and point to complaint options.

Complaint handling

Internal complaint system for platform users; access to out-of-court dispute settlement.

Transparency reports

Regular reports on moderation measures, notices and automated decisions.

Special rules for marketplaces and advertising

Online marketplaces must identify traders before admission (know your business customer), verify the information and inform consumers when they learn of illegal offers. For advertising: ads must be recognisable as such, including the advertiser and the main parameters of targeting. Profiling-based advertising to minors and based on sensitive data is prohibited. Dark patterns in design are banned.

Where AI eases implementation

Many DSA obligations are high-volume processes: receiving notices, reviewing content, giving reasons for decisions, producing reports. AI reduces the effort significantly here — under two conditions: people keep the decision in contentious cases, and the use of automated procedures is disclosed in the terms and transparency reports, as the DSA requires.

  • Pre-classification of reported content and prioritisation by urgency
  • Drafts of statements of reasons referencing the rule violated
  • Pattern detection (repeat offenders, coordinated notices)
  • Automated compilation of transparency reports from moderation data

How we support DSA implementation

We classify your service, check existing processes and texts against the obligations and design notice, reasoning and complaint processes so they work day to day — including the AI-supported steps and their documentation. Legal assessments are made together with your law firm; we deliver the procedural and technical implementation.

privacy first ai

Thinking DSA and AI Act together

Anyone who moderates content with AI support or lets users interact with AI systems touches both frameworks: the DSA's transparency obligations and the EU AI Act's labelling and literacy obligations. We implement both in one process.

Classification

Service category under the DSA and risk class under the AI Act per feature — documented and justified.

Processes

Notice, reasoning and complaint procedures with clear roles, deadlines and logging.

Transparency

Disclosure of automated procedures in terms, labels and reports.

process

How DSA implementation runs

Four steps from classification to ongoing operation.

  1. Step 1: Classification & gap analysis

    Categorisation of the service, comparison of existing processes and texts with the obligations.

  2. Step 2: Process design

    Design of contact points, notice and complaint procedures, reasoning templates and report structure.

  3. Step 3: Implementation with AI

    Introduction of tools for pre-classification, statements of reasons and reports — with human in the loop.

  4. Step 4: Operation & reports

    Regular operation, KPIs, production of transparency reports, adaptation to new guidelines.

contact

Let's talk

Happy to help with anything AI. We will get back to you promptly.

AddressCarolastraße 4-6, 09111 Chemnitz

Get in touch