8 Support Metrics We Monitor for Agentforce on Salesforce Help

You’ve invested time and money into your service organization. But are you getting the return you expect? Are AI innovations like Agentforce actually improving your customer experience?

For more than a year, Agentforce on Salesforce Help has been giving customers faster answers and expert guidance when they need it, without making them search through pages of content or wait on a support ticket. To understand what’s working, we track key support metrics and use those insights to continuously improve the experience..

And here is the honest part. These metrics are evolving. After a full year of operating Agentforce at scale, we realized we needed a new metric to better understand customer intent and create even more transparency into how customers engage. We are learning right alongside our customers. When we discover a better way to measure performance, we adopt it and share it.

We’re sharing the support metrics we monitor and best practices we use to evaluate Agentforce’s effectiveness. These are the metrics that matter most to us. Use them as a starting point and tailor them to your business goals.

What does success look like?

We group our support metrics for the Help Agent into two categories: Adoption and Effectiveness.

#1. Adoption

Before Agentforce can be effective, customers need to use it. Adoption shows whether customers are turning to the Help Agent and gives us early indicators of reduced case volume and cost savings.

We look at:

• Help Portal Sessions: How many customers visit Salesforce Help.
• Conversations: The number of times customers initiate a conversation by submitting a question.

When someone arrives on Help with a question, Agentforce uses natural language understanding to grasp the intent and deliver an accurate, fast response. A strong first interaction builds trust, and customers are more likely to return the next time they need help or want to learn a best practice. When Agentforce efficiently handles common inquiries, our support engineers can focus on deeper or more complex cases.

Tracking adoption helps create positive experiences and build a more scalable service operation.

#2. Effectiveness

Once adoption is established, the next question is the most important: Is Agentforce helping customers?

Customer interactions typically fall into three buckets: resolution by Agentforce, a hand-off to a human support engineer, or a customer abandoning the interaction.

To understand how well Agentforce performs in each area, we monitor:

• Abandons: Customer starts a session but leaves before asking a question.
• Resolutions: Customer interaction concluded without a handoff to a human.
• Customer-Confirmed Resolution: Customer confirms that Agentforce solved their issue.
• CSAT: Customer provides customer satisfaction score at end of Agentforce interaction. 
• Hand-Offs to Human (Escalations): Customer is transferred from Agentforce to a support engineer.
• New: Immediate Customer Initiated Hand-Off

To add more clarity within ‘hand-offs to human’, we introduced a new metric this year. A small percentage of customers start a session intending to talk to a human right away. From day one, one of our core design principles has been to make it easy to reach a human, and we continue to honor that choice.

By tracking these interactions separately, we get a clearer picture of customer intent. This helps us distinguish between sessions that route to humans because a customer requested human help versus sessions that route to humans because Agentforce could not resolve the question. It also shows how many customers are simply using Agentforce as the front door to create a case.

A real-world example

Imagine hitting an integration error that stops your workflow. Instead of hunting through documentation or waiting for a support queue, you ask Agentforce. Within minutes, it identifies the issue and walks you through the fix. If you do need a human, it passes all the context along so you don’t repeat yourself.

Together, these support metrics show how Agentforce contributes to faster resolutions, higher customer satisfaction, and stronger trust. 

Keep tabs with a support metrics dashboard

To understand how Agentforce is doing and whether it is meeting customer expectations, we use a dashboard to track all the metrics.

As we review the data, we ask:

• What’s getting better? Increasing engagement rates, improving effectiveness, and more positive customer outcomes indicate overall progress.
• Where are people dropping off? If customers are abandoning conversations or experiencing delays, it may indicate customer adoption issues or that Agentforce is not providing relevant answers.
• Where are they getting the best answers? Interactions leading to successful resolutions and high satisfaction indicate positive customer experiences.
• What is Agentforce unable to resolve? Understanding these gaps helps us pinpoint areas for improvement so we can refine Agentforce’s knowledge and enhance its ability to provide helpful answers in the future.

A dashboard gives us clarity into how the Help Agent is performing and where we should focus next.

We share our metrics and insights so you can see exactly what we are doing and learn from our approach. Trust comes from transparency. If you visit Salesforce Help, you will see at the bottom how many Agentforce conversations we have handled, how many routed to humans, and more.

Start measuring to drive improvements

Getting Agentforce right requires more than deploying technology. It takes understanding how customers engage, how Agentforce performs, and whether you are achieving results.

• Track the right support metrics for your goals
• Use a dashboard to stay organized
• Be transparent and keep improving

Want to dive deeper into using data and AI for smarter decisions? Check out Critical Thinking and Decision-Making with Data and AI on Trailhead.