I am a Quality Management supervisor and I am currently evaluating over two hundred interactions per week. We have recently implemented Agent Assist to help our agents with complex technical queries. I require a report that correlates the use of Agent Assist suggestions with a reduction in Average Handling Time. However, I have reviewed the standard analytics views and I cannot find a specific metric that indicates when an agent actually clicked or used a suggestion provided by the AI. Is there a hidden participant attribute or a specific event in the conversation detail records that I can use to identify these interactions for my report?
I am currently evaluating Agent Assist for our strategic roadmap and I have found the exact same gap in the reporting! It is incredibly frustrating that the system provides these ‘Intelligent’ suggestions but does not provide a native way to track their efficacy in the standard dashboards. Ership without this data! I have been told that we might have to build a custom event listener that triggers every time an agent interacts with the assist widget, but that seems like a lot of overhead for something that should be built-in.
Greetings! I am so excited to see people pushing the boundaries of Agent Assist! We actually solved this exact reporting challenge for one of our large-scale migrations! You are correct that it is not a standard metric in the Performance views. However, you can find this data by querying the /api/v2/analytics/conversations/details/query endpoint. When an agent uses a suggestion, the platform logs a specific ‘Intent’ or ‘Knowledge Search’ event within the agent’s participant segment.
We built a Java-based integration that parses these detail records and pushes the ‘Agent Assist Usage’ count into a custom field in our reporting warehouse. It is a brilliant way to finally see that ROI you are looking for!