The ability to use existing ticket data as a resource would provide a lot of value. Additional data sources could add value as well. What about documentation tools (IT Glue/Hudu) or communications platforms like Teams? This data could provide value in helping to solve new problems as well. My team regularly assists each other in solving problems in Teams and documents fixes in Hudu. This is a great idea!

On the surface this sounds great but I fear it requires a large amount of work on the part of the MSP as well as the Agent.

How do you identify quality in resoltutions? I know that in ticket there are often many questions and answers which are not high quality solutions. do we go through 10,000 tickets only select the ones that rate as 'high quality'?

do we create a secondary database of quality responses from the original ticket data?

Do we use just the question and answer or the enitre body of the ticket.

Do we run a test/training session to rate results to the agent to train it?

As with any machine learning/rag model, output is directly related to the quality or input and feedback.

curious to see where this goes.