Tropokolagen.pl: an AI assistant on a medical brand website

A configured assistant, together with a knowledge base, a contact path and a dashboard for the team.

• Chat widget - an assistant built on Prose Captain, embedded in the Tropokolagen.pl site.
• Knowledge base - the brand's documents and materials as the basis for answers consistent with the company's official messaging.
• Contextual logic - answers linked to the page the user is on, and navigation across the site structure.
• Start screen - FAQ questions managed from the dashboard and adjusted to campaigns, the season or the most common queries.
• Contact path - detecting contact intent, a form with a phone number or email address, and notifying the team of the enquiry.
• Dashboard - number of sessions and messages, answer accuracy, sentiment, top topics and the moments users move to the form.

Tropokolagen.pl products are complex by nature. Customers wanted to quickly understand the differences between variants, how they work, what they're for and which product is right for them.

The information was on the site as descriptions, materials and documents, but for many users it was scattered or took time to find. These journeys often ended with users leaving the site.

What was needed was a tool that makes the offering easier to understand, helps with the decision and works as a guide to the brand's products.

The assistant combines three layers of information: knowledge from the brand's documents, the content of the page the user is on, and the anonymised conversation context.

With access to the site structure, the assistant doesn't stop at a text answer: it points to specific pages, sections and materials. Conversations begin on a start screen with questions the client's team manages on its own.

From conversation to enquiry

When contact intent comes up in the conversation, the assistant suggests a form with a phone number or email address. Once it's filled in, the enquiry goes to the client's system and the team is notified of the new contact.

Users get fast, consistent answers without searching the site, and the chat has become a channel for generating enquiries.

The company also gained a source of data on customers' real needs, which had previously been scattered or invisible. Structured information on conversation topics and answer accuracy helps improve both messaging and the offering.