Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Return on investment (ROI) and improved customer satisfaction (CSAT) are top priorities for the organizations that implement Copilot Studio agents.
Optimizing the agent deflection rate is one of the top focus areas for organizations to achieve their business goals around ROI and CSAT, and to improve overall agent performance. Major indicators in Copilot Studio that help improve agent performance include resolution rate, escalation rate, and CSAT.
While the metrics continue to evolve, you can take several steps as a maker to improve the deflection rate of your agent. These articles cover the importance of deflection in conversational AI and general techniques and considerations that are universal for optimizing deflection for agents.
Tip
In the context of conversational AI, deflection is an indicator representing the percentage of requests that are completed in a self-service fashion that live representatives would otherwise handle. In other words, it refers to the number of items a team avoids having to deal with as a result of automation.
Why deflection optimization?
Optimizing deflection rates in your Copilot Studio agents provides clear benefits:
- Better customer experience: More customers or employees can resolve their issues with the agent instead of waiting for a human representative in chat or on the phone. This benefit improves the customer experience and increases CSAT scores. While this benefit helps reduce wait time, live representatives can also focus on more complex, higher-value tasks.
- Cost savings: Use deflection rates to measure a key part of the agent's ROI. Human representative call support typically costs around $5 to $10 in the contact center industry, while an agent session that resolves a customer request costs about 50 cents. This cost difference means that higher deflection rates lead to higher cost savings.
Understanding the key components of Copilot Studio analytics to improve deflection
To determine what deflection means for your organization, you need a basic understanding of the available analytics. The following table describes the key metrics from Copilot Studio:
| Description | Details |
|---|---|
| Total Sessions | The total number of analytics sessions within the specified time period. A conversation with an agent can generate one or multiple analytics sessions, each with their own engagement status and outcome. An analytics session begins when a user has new questions after an initial conversation completed (for example, reached the End Of Conversation topic). |
| Engagement Rate | The percentage of total sessions that are engaged sessions. An engaged session is an analytics session where a custom topic is triggered, or the session ends in escalation. Engaged sessions can have one of three outcomes—they're either resolved, escalated, or abandoned. |
| Resolution Rate | The percentage of engaged sessions that are resolved. A resolved session is an engaged session where the user receives an End Of Conversation question that asks "Did that answer your question?" and the user either doesn't respond or responds "Yes." |
| Escalation Rate | The percentage of engaged sessions that are escalated. An escalated session is an engaged session escalated to a human representative. |
| Abandon Rate | The percentage of engaged sessions that are abandoned. An abandoned session is an engaged session that isn't resolved or escalated after one hour from the beginning of the session. |
| CSAT | The graphical view of the average of CSAT scores for sessions where customers respond to an End of Conversation request to take the survey. |
You need to continuously improve these metrics to optimize the agent ROI. Each organization defines deflection rate differently. For example, one organization might consider the Abandon rate along with the Escalation rate as part of their deflection calculation, while another organization might look purely at the Escalation rate.
Although definitions of deflection rate may differ, these metrics form the basis for calculating deflection. Based on our experience with various customers, Resolution rate, and Escalation rate play a major role in case deflection. Increasing the Resolution rate and reducing the Escalation rate typically directly improve overall agent deflection metrics.
This diagram illustrates key techniques to optimize deflection rates in your Copilot Studio agents:
Next step
Learn how to analyze escalation patterns by using built‑in and custom analytics so you can identify high‑impact topics, understand why users escalate, and make targeted improvements that increase your agent's deflection rate.