Agentforce is Salesforce's biggest bet since the cloud transition. The consumption-based $2/conversation pricing could unlock massive scale if enterprises deploy agents widely. But the AI agent market is nascent and intensely competitive -- Microsoft, ServiceNow, Google, and open-source frameworks all offer alternatives. The key question: will Agentforce's CRM data advantage create a defensible moat, or will agents commoditize?
Will $2/conversation pricing scale, or will enterprises demand volume discounts that compress margins?
The unit economics of Agentforce are compelling at face value: 85-90% cost reduction vs human agents. The key variables are resolution rate (currently 30-50% for Tier 1), inference cost trajectory (declining ~50%/year), and whether enterprises adopt broadly or keep agents limited to simple use cases.
The AI agent market is a gold rush with every major tech company competing. Salesforce's advantages are CRM data, customer relationships, and enterprise security. But the risk of commoditization is real -- if agents are just wrappers around LLM APIs, the moat is thin.