Anyone registering a trade mark often faces the same challenge: determining the appropriate classes of goods and services. The so-called Nice classification is the basis for this. It ensures that each trade mark is clearly categorised - an important step in defining the scope of protection and comparability.
However, manual searches in the PDF lists of the German Patent and Trade Mark Office (DPMA) are time-consuming and error-prone. This process can be significantly simplified with modern AI tools.
How classification with AI works
Download Nice classes
On the website of the DPMA you will find the current Nice Classification (12th edition, version 2025) - including alphabetical lists and explanatory notes.
Download the three central PDF files:Instructions & class title
Alphabetical list of goods
Alphabetical list of services
Upload files to an AI tool
Then use any Large Language Model (e.g. Perplexity.ai, ChatGPT or Claude).
After a free registration you can upload the downloaded DPMA PDFs.
Use the correct prompt
With a clearly defined prompt, you give the AI the framework to work exclusively on the basis of the official DPMA documents.
Example prompt
Conclusion
This method is used to determine the correct Nice classes Simple, comprehensible and consistent - even without prior legal knowledge.
AI does not replace the technical review, but it helps to systematically pre-structure and save time.
Florenus advises companies holistically on IP management issues with a focus on economic impact instead of purely legal considerations.
If you want to strategically align your IP portfolio, AI-supported processes can help you operate much more efficiently.
Video instructions
Use AI to determine your brand classes in minutes instead of hours.
EN 
DE 