Trying to understand the calculation logic for Bot Deflection Rate within the Performance Dashboard. Our custom Architect flow logs a 12% resolution rate via Data Actions, yet the dashboard reports only 8% for the same period. The documentation (Genesys Docs) states that deflection requires explicit intent matching, but our logs confirm successful resolution. Are there specific filtering criteria applied to the dashboard view that exclude certain interaction types?
You should probably look at at the specific definition of “deflection” within the Analytics API versus what your Architect flow logs as a successful resolution. There is often a disconnect between custom Data Action outcomes and the platform’s native metric calculations.
The Performance Dashboard relies on the /api/v2/analytics/conversations/summaries endpoint, which aggregates data based on explicit intent match events rather than arbitrary flow completion. If your bot resolves an issue via a fallback mechanism or a custom variable assignment without triggering the final intentMatch event in the conversation metadata, the system will not count it as a deflection.
Check your flow logic. Ensure that the final node before ending the conversation explicitly sets the intent match status. Additionally, verify if any interactions are being filtered out due to session timeouts or if the bot is transitioning to a human agent before the deflection flag is committed. The 4% gap likely represents conversations where the resolution occurred, but the technical criteria for the dashboard metric were not met.
This is typically caused by the difference between raw flow execution logs and the aggregated metrics used in the Performance Dashboard. Genesys Cloud calculates deflection based on explicit intent matches and specific outcome tags, whereas custom Data Actions might log a resolution without triggering the required metric event. This distinction is crucial for accurate reporting.
Coming from Zendesk, where a solved ticket status is the ultimate truth, this asynchronous metric calculation can feel opaque. In Genesys Cloud, you must ensure your Architect flow explicitly sets the outcome to resolved or deflected via a Data Action. Without this explicit tag, the Analytics API ignores the interaction for deflection calculations.
Check your flow configuration to ensure the Data Action includes the correct outcome parameter. The endpoint /api/v2/analytics/conversations/summaries relies on these tags. Aligning your flow logic with the platform’s metric definitions will bridge the gap between your 12% log rate and the 8% dashboard figure. This alignment is essential for accurate migration validation.