Genesys Cloud - Main

We have implemented Copilot in our clients, but in messaging interactions, it's common for customers to send fragmented messages like: 'hi', 'I need help', 'I made a purchase'. But the Copilot doesn't understand context very well with messages like this; it updates with each new customer message and often provides an article that is not in context with what the customer is stating.

Are there any tips to improve the suggestions in this scenario?

Hello Debora,

That's a good question, I am sure the community will be able to help with this.  

Have you reviewed the Resource Center article: Best practices to build and test your natural language understanding

It has some useful tips 

We have implemented Copilot in our clients, but in messaging interactions, it's common for customers to send fragmented messages like: 'hi', 'I need help', 'I made a purchase'. But the Copilot doesn't understand context very well with messages like this; it updates with each new customer message and often provides an article that is not in context with what the customer is stating.

Are there any tips to improve the suggestions in this scenario?

Hi Samuel,

Yes, we've been following these best pratices, but when customer send fragmented messages, copilot cannot understand the hole context and constantly agents receives non related articles.

It would be nice if it had the 'memory' feature like others tools do.

Hello Debora,

That's a good question, I am sure the community will be able to help with this.  

Have you reviewed the Resource Center article: Best practices to build and test your natural language understanding

It has some useful tips 

We have implemented Copilot in our clients, but in messaging interactions, it's common for customers to send fragmented messages like: 'hi', 'I need help', 'I made a purchase'. But the Copilot doesn't understand context very well with messages like this; it updates with each new customer message and often provides an article that is not in context with what the customer is stating.

Are there any tips to improve the suggestions in this scenario?

We have implemented Copilot in our clients, but in messaging interactions, it's common for customers to send fragmented messages like: 'hi', 'I need help', 'I made a purchase'. But the Copilot doesn't understand context very well with messages like this; it updates with each new customer message and often provides an article that is not in context with what the customer is stating.

Are there any tips to improve the suggestions in this scenario?