Not long ago, when we talked about bringing AI into Salesforce, the conversation was fairly simple: what can we automate?
Today the question has changed.
An agent can now analyze information, make decisions within certain limits, take action and pursue a goal for much longer than a single conversation.
Which raises a far more uncomfortable question:
What are we actually willing to delegate to an AI?
- Automating is not delegating. An automation follows rules. An agent can interpret context and decide which action to take. That changes the game.
- The problem won't be technology. It will be governance. What can it do on its own? What needs human approval? What information can it access? How do we know why it made a given decision? Salesforce is putting a lot of focus precisely on autonomy with controls, context and traceability.
- Before implementing Agentforce, look at your processes. Putting an agent on top of a messy process doesn't transform the business. It automates the mess.
Salesforce's next leap won't just be about having better agents.
It will be about the companies that can answer three questions: what to delegate, what to control and what should never be left without human intervention.
That's where the real conversation about AI Governance begins.