Past conversations

Discover how glacom.AI manages dialogue memory and how this information helps improve services.

How it works

Your assistant remembers what happens during the current chat to keep the flow of the conversation. For example, if a customer asks "What are your hours?" and immediately after writes "Are you open on Saturday?", the bot understands they are still referring to the opening hours.

However, concluded chats do not automatically become knowledge sources. If a customer mistakenly writes that your product costs 10 euros instead of 100, the bot will not learn this wrong data to answer other users. This prevents the system from absorbing errors or sensitive personal data without your control.

What you need to do

Check the conversation history periodically. If you notice many customers asking a question the bot cannot answer, you can use that information to update the sources or follow the guide on how to add what is missing.

What glacom® does

glacom® keeps the history of all interactions. This data is used for controlled training: we analyze where the bot was imprecise and refine the answers to make it more effective over time, following the path described in training.

Limits

The bot does not have automatic "long term memory" between different sessions. If a customer returns after a month, the bot will not remember the details of the previous conversation unless they were saved via integrations.

Frequent problems

• Did the bot forget what I said two days ago? → Yes, every new session starts from scratch to ensure privacy and accuracy.
• Can I turn a chat into a source? → Not automatically: you must extract the correct information and add it to the external data.
• Does the bot learn from customer mistakes? → No, learning only happens through human supervision.

From the app

You can consult the conversation history and monitor dialogues in real time.