Zendesk wants to provide you with the insights needed to be successful with Generative AI Agents (also known as chat bots). The automation potential page will highlight the areas where you are well equipped to offer a generative answer to your customers, provide metrics to highlight the potential benefits, and help you identify knowledge gaps which you may decide to accommodate so that your AI agents can answer more topics.
Through the use of AI analysis, we are able to extract common topics in your customer conversations and allow you to understand the efficacy of the generative AI agent in being able to serve those conversations. By automating these conversations, you will be able to free up your agents to work on more complex or unique issues.
To read more about this feature, please see Viewing and using the automation potential report to create or enhance AI agents.
Understanding your Automation Potential
Accessing Automation Potential
To see your Automation Potential dashboard, you can go to Admin Center > AI > AI Agents and click the ‘View Automation Potential’ button in the upper right corner.
Automation Potential organization
The insights are segmented by brand. Brands are often used to separate different customer cohorts as well as control which content might be used by the AI Agent. Within each brand's report, you can see the predicted savings and automation potential which was identified from the tickets sampled in your account.
Automation Potential metrics
Once your tickets are analyzed, the volume of one-touch tickets are compared to the total volume of tickets with that subject to understand how much of your volume is likely to be automated. The monetary and time savings are based on a static benchmarked cost of ticket value (we've started with $3.50 USD, but will be happy to discuss how you measure cost of service), and the calculation for time savings is represented by 10 minutes per ticket as that is forecasted as what could be saved across those one-touch tickets and the average time it takes an agent to compose a reply to common tickets. These definitions are likely to evolve and are a topic for which we will certainly be gathering your feedback.
Conversation insights
Conversations ready to automate
Once your tickets have been categorized, you will see the ability to expand each parent subject to see the topics contained within. With the first release, we identified topics which you already have documented in your help center. This lack of knowledge gap represents a readiness to automate for this topic or question. We also provide some of the sample tickets which were analyzed to raise these insights so you can confirm if the topic and conversation match what your customers are asking.
In our discovery with you, we will also be looking to discuss topics where knowledge was not found and how you might act on these insights.
Within each of these topics, we show the ability to view the sample response. This demonstrates to you how the Zendesk Essential AI Agent would reply to the identified topic, so you can see how it would respond to your customers.
Knowledge gaps
You will see a second tab in the automation potential interface which is labeled for knowledge gaps. This is where you can see topics from your customer conversations where a respective answer was unable to be found. The aim for these topics will allow you to fill the top knowledge gaps in aims to help your AI agent answer more questions.
Any questions about the functionality, feel free to comment below. If you have noticed an issue in your account specifically or want to provide feedback, please report it by following the guidance in this post.
