Genesys Cloud - Main

Hi all - does anyone have experience with AI Insights in their organization?

I conducted a test call to check the AI Scoring for a particular call type.

Note: for this test call, incorrect information was deliberately provided.

In addition, we were not using the knowledge articles for this test. (I'm not sure if or how they are connected)

The Evaluation was scored as expected (satisfied our prompts are informing the AI Scoring). However, I noticed under the Resolution Section of AI Insights that it stated "the agent provided the correct" information and stated the Resolution Status as Resolved.

How does Genesys AI Insights determine resolution status? Does it mark any Q&A as resolved regardless of accuracy? Should we use knowledge articles to help it assess factual correctness?

How can we improve the Insights provided by AI Insights? 

Any tips on optimizing AI Insights and AI Scoring would be appreciated.

You are mixing technologies here.  

AI Insights provides conversation-level summarization along with reason for call, outcome of call, and follow-up items all provided by the AI from analysis of the transcript alone - no other information is provided to the AI.

AI Scoring is and AI-based evaluation form that again looks at only the transcript to answer questions based on the question, the help text, and the answers.  

Knowledge is really only used by Bots and CoPilot to provide answers to questions asked by the caller.

So, your tests seem to be looking at the result of the AI-provided reason, outcome, and follow-up items.  Currently, we have very few controls for these and the summary aside from the transcription, which you can affect using the Dictionary Management. 

While I don't know of any changes to this in the near-term roadmap, Genesys is planning on BYO LLM Q1 '26 that should allow you to pump the transcript through your own LLM that could be programmed to manipulate this output.  I'll see what is planned for the "in-skin" LLM. 

Hi all - does anyone have experience with AI Insights in their organization?

I conducted a test call to check the AI Scoring for a particular call type.

Note: for this test call, incorrect information was deliberately provided.

In addition, we were not using the knowledge articles for this test. (I'm not sure if or how they are connected)

The Evaluation was scored as expected (satisfied our prompts are informing the AI Scoring). However, I noticed under the Resolution Section of AI Insights that it stated "the agent provided the correct" information and stated the Resolution Status as Resolved.

How does Genesys AI Insights determine resolution status? Does it mark any Q&A as resolved regardless of accuracy? Should we use knowledge articles to help it assess factual correctness?

How can we improve the Insights provided by AI Insights? 

Any tips on optimizing AI Insights and AI Scoring would be appreciated.

Hi all - does anyone have experience with AI Insights in their organization?

I conducted a test call to check the AI Scoring for a particular call type.

Note: for this test call, incorrect information was deliberately provided.

In addition, we were not using the knowledge articles for this test. (I'm not sure if or how they are connected)

The Evaluation was scored as expected (satisfied our prompts are informing the AI Scoring). However, I noticed under the Resolution Section of AI Insights that it stated "the agent provided the correct" information and stated the Resolution Status as Resolved.

How does Genesys AI Insights determine resolution status? Does it mark any Q&A as resolved regardless of accuracy? Should we use knowledge articles to help it assess factual correctness?

How can we improve the Insights provided by AI Insights? 

Any tips on optimizing AI Insights and AI Scoring would be appreciated.