When Salesforce launched Agentforce 360 in October 2025, CEO Marc Benioff called it a milestone moment for enterprise AI. A year into its rollout, the platform has been adopted by roughly 12,000 customers — including companies like Reddit, OpenTable, and staffing giant Adecco — and it's become one of the clearest real-world examples of what "agentic AI" actually looks like inside a business, as opposed to a slide at a tech conference.

What Agentforce 360 actually does

The platform brings together several previously separate Salesforce products — Data 360, Customer 360 Apps, and Slack — into one system where AI agents can act with business context rather than operating as a standalone chatbot. Its components include the Atlas Reasoning Engine, which handles the agent's decision-making, an Enterprise MCP Registry that lets agents connect to outside tools and data sources, and features like Intelligent Context that let an agent pull in relevant account history or case data before acting. In practice, this means an agent can resolve a customer support ticket, update a CRM record, and flag an account for a human rep to follow up on, all in one pass, instead of requiring a human to do each step manually.

Why this matters for sales and support careers specifically

Sales operations and customer support are the two functions where Agentforce's agents are most directly doing work that used to sit squarely with entry-level and mid-level human staff: first-line ticket triage, routine order status queries, basic lead qualification, and account data entry. Salesforce has positioned Slack as the human interface into this system — the idea being that employees don't talk to a separate "AI tool," they work alongside agents inside the same channels they already use.

What's actually changing on the ground

• Tier-1 support roles are shifting toward exception handling. Where AI agents can resolve the routine volume, human support staff increasingly spend their time on the harder cases the agent escalates — which raises the skill bar for what counts as an "entry-level" support job.
• Sales development roles are shifting toward judgment and relationship work. Lead qualification and initial outreach drafting are increasingly agent-assisted or agent-led, pushing human SDRs toward the parts of the funnel that require reading a prospect's intent or handling objections — skills that are harder to template.
• A new skill category has opened up: agent oversight and configuration. Someone has to design the workflows these agents follow, set guardrails on what they're allowed to do autonomously, and audit their decisions. Salesforce's own partner ecosystem — over 500 pre-built features available to partners — has created demand for people who understand both the business process and how to configure an agent to follow it.

What to do with this if you're in sales or support today

If your current role is heavily weighted toward the routine, repeatable parts of sales or support work, the near-term risk isn't that your job disappears overnight — Salesforce's own case studies lean on agents handling volume alongside, not instead of, human teams. The more realistic risk is that the routine portion of the role shrinks, and the bar for what counts as doing the job well rises to include judgment calls, escalation handling, and increasingly, the ability to work with or configure the agents themselves. Building familiarity with how these systems are set up — even at a basic, practical level — is quickly becoming as relevant to a support or sales career as traditional CRM skills were a decade ago.