Agentforce Command Center lets admins monitor and fix deployed AI agents without code. The demo shows how sales agents track revenue and deal cycles while service agents measure customer satisfaction and case costs, then walks through diagnosing a bad cancellation interaction and fixing it with a natural-language instruction in AgentForce Builder.
Questions this video answers
June 24, 2025
agentforce.com
Chapters
In this demo, I'll show you how AgentForce Command center helps you supervise and improve your AI agents. This is AgentForce Studio, and this is where you'll build, test and manage your AI agents. And what you're looking at here is a list of the agents I've built and deployed with AgentForce. But once you've deployed your first agent, you'll have questions about how the agent is performing and how to make it better. And that's why soon, Command center will become part of AgentForce Studio. Command center is where you'll supervise your AI agents. For example, what you're looking at here are the metrics for our sales agents, which help our sales teams qualify leads and close deals. And notice how these metrics are adapted to that type of agent. We're seeing how AgentForce is impacting our sales revenue, our deal cycles, and our team's productivity. But watch this. If I select our service agents, notice how all these metrics are different. Now we're seeing how Agent Force is impacting our customer satisfaction, our case deflection, and the average cost of our case resolutions. And this is key because different types of agents have different criteria for success. And that success can only be understood in the context of your business Goals. And because AgentForce is part of the Salesforce platform and your customer360, it's the only platform that has visibility into your business. But Command center does more than help you understand the impact AgentForce is having on your bottom line. It also helps you find opportunities for improvement. For example, here we're seeing low quality scores for cancellation requests. And if we click into this item now, we're starting to dive into the details and we're seeing the individual interactions that are contributing to these low scores. And with Command center, you can zoom all the way into an individual interaction and see exactly what is happening. Here we're seeing a customer struggling to change the dates on their reservation, which sends the customer down the cancellation path when what the customer really wanted was a resolution to their core issue. And with this insight into what is going wrong inside Agent 4 studio, you can make improvements. So here we are in Agent 4's builder, where admins and developers can create and improve their agents. So to fix this issue, I'm going to add a natural language instruction, not code, that instructs AgentForce to only use this cancellation topic if explicitly asked for a cancellation. Then here inside agentforce Builder, I can test this to ensure it's working as expected and no longer sending customers down that cancellation path. Then AgentForce could help me batch test and when I'm ready, activate and deploy. And after deployment, Command center is still helping you supervise your agents so so you can validate these improvements as you constantly convert those challenges into the desired business outcomes. So now you've seen how AgentForce Command center helps you supervise and improve your AI agents. To learn more, visit agentforce.com thanks for watching.
