Generative AI

Generative AI writes rather than just sorts. Where earlier methods assigned a request to a category, a language model drafts a reply, summarises a conversation or translates a message.

#How language models work

A language model has been trained on very large amounts of text. It calculates which wording is likely to follow an input. The result is text that reads fluently and sounds plausible.

Plausible does not mean checked. The model knows neither your tariffs nor your deadlines unless you provide them as a source. Without that link, hallucinations follow.

#Where it helps in customer service

• Summarise. After the conversation, a note with request, outcome and next step is ready.
• Draft. Suggested replies for email and chat that agents check and adapt.
• Translate. Answer messages in the customer's language.
• Simplify. Turn technical text into plain language.
• Answer from knowledge. Respond with content from the knowledge base, with its source.

Univerx Assist· active

Conversation

AI agent, disclosedHandover from AI agent · context complete

I received the decision letter, but the amount does not match my application.

I have the case in front of me. One moment, I am checking the calculation.

Sentiment: calmVoice

Univerx Assist

Suggested reply

The letter deviates from the application. I will propose a correction and send you the confirmation by message.

InsertAccept

Summary

Request: deviation in the letter. Documents on file. Next step: correction, confirmation by message.

#What changes for agents

Generative AI mainly takes writing work off people's hands. Wrap-up shrinks when the summary is already there. The time saved goes to the next conversation, or to a request that needs more attention.

What matters is who has the final say. A suggestion stays a suggestion: agents accept, edit or discard it. That keeps responsibility for the answer where it belongs.

#Fit and safeguards

#Generative AI in Univerx

In Univerx, generative features sit in several modules: Univerx Assist provides suggested replies and summaries, and Univerx Knowledge answers questions from your own sources. Generative answers rely on your knowledge base, not on the model's general world knowledge.