I am a reporting analyst and I am completely fed up with the lack of granularity in the standard bot performance dashboards! We have a very complex bot architecture that uses several ‘Sub-flows’ for modularity. The problem is that the Genesys Cloud analytics engine rolls up all the sub-flow interactions into the parent flow’s metrics. I cannot see which specific sub-flow is failing or which intent recognition issues are happening deep inside the sub-modules! It makes my executive dashboards look like a total guess. Why is there no native way to drill down into sub-flow performance, or am I just missing a hidden filter in the Architect Flow views?
Hello. I understand your frustration. We use Terraform and CX as Code to manage these complex flows, and we ran into the exact same reporting gap. The platform treats the entire bot session as a single execution ID. To get the granularity you need, you must use ‘Flow Outcomes’ or ‘External Tags’ at the start and end of each sub-flow. In our Terraform scripts, we automatically inject these tags into every sub-module. This allows you to filter your analytics queries by the tag, giving you a clear view of where the customer entered and exited each sub-flow.
Greetings! Transitioning from PureConnect to Genesys Cloud has shown me many reporting differences. In PureConnect, we had much more control over handler-level tracing. To achieve something similar here, you should consider utilizing ‘Participant Data’ to log the name of each sub-flow as the customer traverses the bot. While it does not show up in the standard ‘Bot Performance’ view, you can pull this data into a custom ‘Interaction Detail’ report. It is a very helpful methodology for identifying those intent recognition failures that the executive team is worried about.
As a Senior Architect, I can confirm that the standard dashboards are definitely limited for nested architectures. The most robust enterprise solution is to query the /api/v2/analytics/botflows/sessions endpoint directly via the API. This endpoint provides the raw event stream for the bot session, including every sub-flow transition and every intent match attempt. We built a custom dashboard for a ConvergeOne client that parses this event stream in real-time. It provides the exact ‘Drill Down’ capability you are looking for without requiring you to manually tag every sub-module in Architect.