Putting AI into production for two decades.

We brought an analytics platform's features to AI assistants such as Claude and ChatGPT through an MCP server, including analytics capabilities that weren't previously available in the product's own UI. An MCP App extension renders interactive charts right in the conversation, so users can filter and drill into the data without leaving the chat.

Users ask about their own data and the product's features in plain language. The assistant retrieves answers from a knowledge base with source citations, suggests follow-up questions and builds charts right in the conversation.

Users describe the view they want, and the model turns the request into a chart specification. The model never generates the numbers itself. All data comes from the source system, so the results stay accurate. Works in multiple languages.

When a new data source is connected, AI recognises its task statuses and maps them to workflow stages automatically. Users can accept or correct the suggestion, and any new statuses are mapped automatically from then on.

A proprietary natural-language ontology with more than 200,000 topics, used for content classification, recommendations, semantic search and contextual targeting in multiple languages across Europe and North America. The solution became part of an international ad-tech company.