Salesforce Agentforce
AI / LLMEnterpriseVisit Site
Salesforce native agent platform, priced per action via Flex Credits — the default choice if your system of record is already Salesforce.
Overview
Agentforce wins on data gravity, not on model quality. If your customers, cases, and pipeline already live in Salesforce, an agent that reads and writes them natively skips the integration work that sinks most agent projects. The pricing has settled into something legible after a rocky start: Flex Credits at $500 per 100,000 credits, where a standard action costs 20 credits ($0.10) and a voice action 30 credits ($0.15). One action covers up to 10,000 tokens — cross that and you are billed for two actions, which is the detail that wrecks naive cost models. There is also a $2-per-conversation option, which is cheaper only if your average conversation exceeds 20 actions. Salesforce Foundations includes 200,000 Flex Credits free on Enterprise Edition and above, so you can pilot before committing. Model your cost on real transcripts, not on averages.
Pros & Cons
Pros
- Native access to Salesforce data with no integration layer
- Transparent per-action pricing after the Flex Credits change
- Free 200,000-credit allocation to pilot on Enterprise Edition and above
- Two pricing models so you can pick the cheaper one for your usage shape
- Inherits existing Salesforce permissions and audit trail
Cons
- The 10,000-token-per-action ceiling silently doubles costs on long actions
- Only makes sense if Salesforce is genuinely your system of record
- Costs scale with usage in a way seat-based buyers find hard to forecast
- Capability trails specialist CX vendors like Sierra and Decagon
Workflows that use Salesforce Agentforce
Get a new AI workflow each week — many feature Salesforce Agentforce and other tools in this category.