Salesforce is betting that Agentforce, its autonomous AI agent platform, has crossed from novelty to revenue line item, and Wednesday’s Q4 earnings call is where that bet gets priced. The company last disclosed 9,500 paying Agentforce customers at $540 million in ARR, with KeyBanc projecting a path to $1 billion. Average ARR per customer has already doubled in six months to roughly $57,000. Six acquisitions in the past quarter, including Informatica’s $8 billion data management business, are being assembled into what Salesforce is positioning as an AI data moat<\/a> competitors can’t quickly replicate.<\/p>\n The question that actually matters for revenue leaders isn’t whether Agentforce works in a demo. It’s whether the monetization model has stabilized enough to build a multi-year expansion plan around. Right now it hasn’t, quite. Customers are still largely on pilot pricing, use cases are concentrated in simple tasks like auto-populating documents and pulling contract clauses, and Salesforce appears to have quietly retired its 10-cents-per-action pricing tier, suggesting the early economics needed renegotiation. If your team is in that pilot cohort, the repricing pressure is coming.<\/p>\n The displacement fear, that Anthropic or another AI-native vendor could route around Salesforce entirely, is probably overstated in the next 18 months but genuinely underestimated over five years. Wedbush, whose optimism on incumbents like Salesforce is shaped by its sell-side positioning, makes the structurally sound point that CRM systems of record aren’t replaced by layering in a new chat tool. The harder question is whether Agentforce captures the orchestration layer before AI-native competitors build enough enterprise distribution to matter. Salesforce’s six acquisitions in one quarter read less like a product roadmap and more like a land-grab to fill that layer before the window closes.<\/p>\n The Informatica acquisition is the most consequential signal here. Clean, governed data, the kind that tells an AI agent what a customer actually did versus what the CRM was supposed to record, is the constraint that breaks most enterprise AI deployments before they scale. Salesforce now owns that pipeline for its installed base. CROs at companies already deep in the Salesforce stack have a defensible reason to expand rather than diversify. CROs at companies still mid-migration should treat the Informatica integration timeline as a dependency in their AI roadmap, not a footnote.<\/p>\n Watch the ratio of new Agentforce logos to expansion ARR on Wednesday. If expansion is doing the heavy lifting while net-new customer acquisition flatlines, that tells you Agentforce is a retention and upsell play for the existing 150,000-customer base rather than a growth engine pulling new buyers into the platform. That’s not a bad business, but it reframes the vendor conversation: the renewal you’re defending with Salesforce this year is also implicitly a decision about whether to be inside or outside that data moat when the next pricing cycle hits.<\/p>\nWhat this means for your business<\/h2>\n