AI & Automation KI and NovaDB
NovaDB integrates AI directly into product data and content processes. From simple automation to autonomous agents within the workflow. AI is an integral part of the platform—not an add-on, and not a vendor lock-in. This enables intelligent workflows that reduce manual effort and significantly shorten time-to-market.
A platform for product data, content, and assets
NovaDB combines PIM, CMS, and DAM into a single integrated system. Product data, editorial content, and digital assets are stored in a shared database and distributed from there to all channels.
This architecture makes NovaDB the ideal foundation for AI: The data model is self-describing, access is API-first, and AI always sees PIM, CMS, and DAM simultaneously. Whether it’s simple text generation or autonomous agents in the workflow, both are available directly within the platform.
NovaDB and AI
Claude + AI Agents + NovaDB: From Structured Data to a Landing Page
In this demo we explore how Claude, AI Agents and NovaDB can work together to transform structured data into web content.
Youtube collects data when loading the video. By clicking on "Load video at Youtube" you agree to this. You can find information in our privacy policy.
From PIM to Published: NovaDB AI Agents in Action
Using AI Agents configured directly within NovaDB, product information stored in the PIM is automatically transformed into fully written blog articles.
Youtube collects data when loading the video. By clicking on "Load video at Youtube" you agree to this. You can find information in our privacy policy.
AI can be used at every level
NovaDB supports AI deployment in two ways. AI functions run on demand: text is generated, a field is translated, or an asset is automatically tagged. Quick to set up, ready to use immediately.
AI agents go a step further. They understand context and data models, independently identify when action is needed, and execute multi-step tasks. They are active participants in the workflow with their own identity, roles, and permissions, and can be addressed directly via @mention. Decisions are always made by the team
Both approaches are available directly in NovaDB. You start with simple AI functions and expand to agents at your own pace.
AI in PIM: How Agents Contribute to Product Data Management
Wenn eine neue Produktserie im PIM angelegt wird, erkennt ein KI-Agent, welche Beschreibungsfelder noch fehlen, welche Variante noch keine technischen Merkmale hat und ob ein ähnliches Produkt bereits als Referenz existiert.
Er erstellt Textentwürfe auf Basis der vorhandenen Attribute, prüft ob Masseinheiten konsistent sind und ob Klassifikationsangaben wie ECLASS oder ETIM vollständig gepflegt wurden. Alle Vorschläge landen als Aufgaben im Workflow. Freigaben bleiben beim Team.
AI in CMS: From Draft Text to Active Workflow Participant
AI agents perform defined tasks without manual intervention: generating product descriptions from PIM attributes, preparing content for various channels, making texts ready for translation, or enriching fields based on rules.
AI agents go a step further. They independently recognize when dependent content needs to be updated following a product change, prepare tasks in the workflow, and can be directly addressed via @mention. Decisions are always made by the team
Control NovaDB directly from AI environments
Using the Model Context Protocol (MCP), NovaDB can be controlled from external AI environments, such as directly from Claude Desktop. What sets this apart from simple data access is that the AI model not only sees product data from the PIM, but also editorial content from the CMS, media from the DAM, and the current workflow status—all based on a shared data foundation.
A product text created in this way knows the associated category text, the available images, and the product’s approval status. The result is not an isolated text suggestion, but context-aware content that fits directly into the ongoing process.
The connection works in both directions. The AI model reads, analyzes, and writes results back directly. Access follows the same roles and permissions as any other use in NovaDB.
Your model, your choice
NovaDB does not dictate a specific AI model. You decide what best suits your needs: privacy requirements, performance, cost, or existing infrastructure.
- OpenAI GPT-4o, GPT-4.1
- Anthropic Claude
- Google Gemini
- xAI Grok
- Private / On-Premise
If you use one model today and switch tomorrow, you don’t have to touch NovaDB. The integration follows open standards without vendor lock-in
Here's how it works in practice
Ask a question
You ask Claude: “Which products in the Pumps category don't have an English description yet?”
AI analyzes and generates
Claude reads the data directly from NovaDB, identifies the gaps, and generates draft text based on the existing PIM attributes.
The result goes into the workflow
The drafts are added as tasks to the editorial workflow in NovaDB. Your team reviews them, approves them, and that's it.
See how AI works with your data in PIM and CMS.
In a personalized demo, we’ll show you what makes sense for your processes today: simple AI features that work right away, or AI agents that integrate into your workflow. You set the pace.
