In today's Cloud Wars Minute, I look at why Salesforce's new Help Agent represents a major shift toward performance-based enterprise AI.
Highlights
00:03 — Salesforce is launching the Agentforce Help Agent, a pre-built AI customer service agent that customers can deploy in a matter of minutes. It's designed as an alternative to custom-built agents, connecting to existing Salesforce knowledge articles and support content, which means only minimal configuration is required.
00:23 — I'm going to walk you through the features of this new agent before getting to the part I'm most excited about, and I think you will be too. The agent was built on the Agentforce platform and uses the Salesforce Data Cloud and CRM data for context, incorporating the responsible use and governance policies there too.
00:43 — It delivers enterprise-grade customer support by answering customer questions, troubleshooting issues, escalating complex cases to a human support agent. Salesforce validated the agent internally before the launch, and the company has reported that its own Help Agent handled 4.3 million customer conversations and resolved around 70% of inquiries autonomously, really showcasing its effectiveness.
01:16 — Here's the kicker: the new Help Agent operates on a resolution-based pricing model. This means that customers are only charged when the agent successfully resolves a customer's issue. There's no charge if the conversation is handed off to a human agent before resolution, and this approach is quite groundbreaking. In many ways, Salesforce is testing a new pricing model for enterprise AI.
01:50 — From an AI in the workplace perspective, this agent operates on a performance-based pricing scenario, similar to how a gig worker is paid for successful tasks completed, right? So, Salesforce is not the first company to use outcome-based pricing for software, but bringing it to Agentforce and pushing it further into enterprise AI is remarkable stuff and a great step forward from Salesforce.
Visit Cloud Wars for more.
