Salesforce raises prices 6% for AI features despite performance gaps

CALIFORNIA, UNITED STATES — Salesforce announced a 6% price increase for its enterprise cloud products, citing enhanced AI capabilities as justification.
The move comes as the customer relationship management (CRM) leader rebrands its AI offerings as Agentforce, with premium plans costing up to $550 per user monthly, while facing customer skepticism about the technology’s reliability.
AI-powered price hikes meet customer skepticism
Salesforce’s latest pricing adjustment affects Sales Cloud, Service Cloud, and select Industries editions, exempting lower-tier plans like Starter and Pro.
The company argues that new AI integrations, including Agentforce add-ons, starting at $125 per user per month—justify the increases, but internal research reveals performance gaps. A Salesforce study found that AI agents succeeded in just 58% of single-step tasks and only 35% of multi-step workflows, undermining confidence in the value proposition.
This time, Wall Street’s reaction has been tepid, with shares quickly retreating after a brief rally. Customer forums reflect growing discontent, with complaints about AI hallucinations and unmet expectations.
As Chief Financial Officer (CFO) Amy Weaver acknowledged, price increases take time to absorb, but with AI’s shortcomings now public, Salesforce may struggle to convince users that the premium is warranted.
Agentforce rebrand aiming to refresh Salesforce’s AI strategy
The shift follows the month of May’s pricing overhaul, replacing per-action fees with credit-based models to offer cost flexibility.
Yet, executives admit that AI pricing remains experimental, with Bill Patterson, Salesforce executive veep for CRM applications, stating that no vendor has figured out the ideal model.
The rebrand extends to Slack, with a new Enterprise+ tier that bundles AI-powered search across platforms.
However, the timing is awkward—as Salesforce promotes these upgrades, its research highlights AI’s unreliability, and security concerns linger after Google reported breaches via fake IT support calls.
For customers, the challenge is weighing promised efficiency gains against the risks of paying more for imperfect tools.