Salesforce launches Agentforce 3, giving firms AI control
CALIFORNIA, UNITED STATES — Salesforce has launched Agentforce 3, a groundbreaking upgrade to its AI-powered digital labor platform that enables seamless collaboration between humans and AI agents, according to their press release.
The new platform has already demonstrated impressive results, including reducing customer case handling times by 15% and autonomously resolving 70% of routine queries for early adopters like 1-800Accountant.
Unprecedented AI workforce visibility
According to Sanjna Parulekar, Senior Vice President (SVP) of Product Marketing at Salesforce, the system allows businesses to track adoption, deflection, and cost in custom dashboards while maintaining clear differentiation between human and AI responsibilities.
In a No Jitter article by Matt Vartabedian, Parulekar notes, “AI agent signals can now surface directly in your contact center wall boards as a way to bring humans and agents into the flow of work.”
The Command Center integrates seamlessly with existing tools like Datadog and Splunk through OpenTelemetry standards, creating a unified operational view.
This level of observability addresses what Ryan Teeples, Chief Technology Officer at 1-800Accountant, calls the critical challenge in AI adoption.
“With a high level of observability, we can see what’s working, optimize in real time, and scale support with confidence,” explains Teeples.
Protocol eliminates integration barriers for rapid deployment
One of the most significant innovations in Agentforce 3 is the Model Context Protocol (MCP). Gary Lerhaupt, Vice President (VP) of Product Architecture, dubbed it a universal translator to AI systems.
“Without MCP, you need to create complicated point-to-point integrations for each new tool, which means you’re handling the API connectivity, authentication data, access code, and more,” Lerhaupt explains, noting MCP’s value.
This innovative technology will render obsolete complex point-to-point integrations that hindered AI adoption in the past, thereby enabling companies to plug agents into off-the-shelf tools as easily as attaching peripherals.
The platform comes with over 100 prebuilt industry actions to accelerate implementation, contributing to explosive growth in adoption. A soon-to-be-released Slack Workforce Index reports AI agent usage up 233% in six months, with 8,000 customers signing up for Agentforce.
Teeples credits the platform with handling 70% of administrative chats during peak tax season, calling it an incredible lift during our busiest period.
Future implications
Agentforce 3 will be a significant step toward developing a truly hybrid workforce, where humans and AI work together efficiently.
The success of the platform is a sign that the corporate world is willing to assign some of its multitasking functions to the AI agents and leave coordinating complex interactions to human skills.
“With Agentforce, we’ve unified agents, data, apps, and metadata to create a digital labor platform, helping thousands of companies realize the promise of agentic AI today,” notes Adam Evans, Executive Vice President and General Manager of Salesforce AI.
