What is Salesforce Agentforce?
Agentforce is Salesforce's platform for building and running autonomous AI agents inside the enterprise. Rather than a single chatbot, it lets companies create agents that act on business tasks, grounded in their own CRM and business data through Salesforce's Data Cloud.
Agents are built with a reasoning engine that plans steps and calls approved actions, and they run inside Salesforce's trust and guardrail layer, so what an agent is allowed to do and see is governed by policy. Common uses are service, sales development, and commerce, where an agent handles routine work and escalates the rest.
It fits organizations already on Salesforce that want governed, data-grounded agents rather than a general assistant. The trade-off is platform commitment and consumption-based cost, so model the economics against the volume of work the agents will handle.
What it's best for
- Companies already on Salesforce that want agents grounded in their CRM data.
- Customer service and sales-development tasks with clear escalation paths to humans.
- Governed deployments where guardrails and data access must follow policy.
- Scaling routine, high-volume interactions without adding headcount linearly.
Where it falls short
- Teams not on the Salesforce platform, where the data-grounding advantage is lost.
- Open-ended general assistance unrelated to business processes and CRM data.
- Buyers who want predictable flat pricing rather than consumption-based cost.
Ways in
Agentforce is configured inside Salesforce, not installed as a standalone app. You define an agent's topics, the actions it can take, and the data it can ground on through Data Cloud.
Start with a narrow, high-volume use case (for example, tier-one support) and a clear escalation rule to a human before expanding scope.
Getting the most out of it
Invest in the data grounding and the action library: an agent is only as good as the trustworthy data and safe actions you give it.
Set guardrails explicitly and monitor outcomes, then widen the agent's autonomy as you build confidence.
What Salesforce Agentforce costs
Approximate, in USD, as of June 2026. Prices change often. Confirm on the official site before you rely on them.
Consumption-based
Per action / per conversation
Priced on usage (for example, per resolved conversation or per action), often via prepaid credits, on top of Salesforce licensing. Confirm current pricing with Salesforce.
Example prompts
Copy these into Salesforce Agentforce as starting points, then adapt them to your task.
Design a support agent that answers order-status questions from our data, issues a return label when eligible, and escalates anything outside policy to a human.
Create an agent that qualifies inbound leads against our criteria, books a meeting when they fit, and logs the outcome to the CRM.
List the data this agent can read and the actions it can take, and confirm each is allowed by our access policy.
Salesforce Agentforce
common questions.
Direct answers to the questions we get asked the most. If yours isn't covered, write to the team.