The excitement surrounding Salesforce's Agentforce exceeds its actual delivery.

      Salesforce has closed 29,000 Agentforce deals and reported $800 million in annual recurring revenue (ARR), yet its stock has dropped by 30 percent in 2026 amid the widespread selloff known as the SaaSpocalypse. Demonstrations from Williams-Sonoma, UChicago Medicine, and SharkNinja were more about potential than actual implementations.

      Salesforce faces a significant issue that no marketing strategy can resolve. The company has centered its narrative around Agentforce, its AI agent platform, presenting impressive statistics: 29,000 deals closed and $800 million in ARR, along with a roadmap suggesting it could replace numerous human jobs. However, Wall Street remains skeptical, with a growing disparity between Salesforce's stage demonstrations and actual customer usage.

      The stock performance reflects this sentiment. Salesforce shares dropped nearly 21 percent in 2025 and have continued to decline another 30 percent in 2026. This downturn aligns with a general selloff of software-as-a-service (SaaS) companies, a phenomenon termed the SaaSpocalypse, where approximately $285 billion in market capitalization was lost within a mere 48 hours in February. The reasoning behind this is straightforward: if a single AI agent can perform the work of ten employees, companies have little incentive to pay for ten software licenses.

      Salesforce has attempted to mitigate this concern by branding itself as a provider of AI agents rather than just software seats, with CEO Marc Benioff referring to Agentforce as a "digital labor platform." During earnings calls, the company often cites the 29,000 deals and ARR as evidence of enterprise engagement.

      However, the showcased examples have not held up under close examination. At Dreamforce, Salesforce demonstrated an AI agent named Olive for Williams-Sonoma, intended to assist customers with meal planning and product selection. In reality, Olive struggled with specific inquiries and recommendations, with advanced features described in future tense rather than as operational.

      A similar situation arose with University of Chicago Medicine, presented as a premier Agentforce for Health deployment. In truth, UChicago Medicine’s initial AI agent only managed basic web chat inquiries like parking directions and clinic availability, with more advanced functionalities still under development.

      SharkNinja, known for its vacuums and kitchen appliances, was another headline client. Salesforce claimed the company would leverage Agentforce to enhance customer service, with Bloomberg reporting a 20 percent reduction in support calls. However, the implementation discussed was also forward-looking, specifying agents would soon "guide customers through the buying process" rather than detailing past achievements.

      This issue is significant as Salesforce is not alone in overstating AI capabilities. In May, Apple agreed to a $250 million settlement over claims that it exaggerated the capabilities of Apple Intelligence and Siri with the launch of the iPhone 16, suggesting their marketing exceeded the functional reality of the technology.

      Salesforce’s financial outlook presents further challenges. Revenue growth has slowed from approximately 25 percent a few years ago to around 10 percent for fiscal 2026, with the company reporting total revenue of $41.5 billion. While still a substantial business that posted a strong fourth quarter with 12 percent growth, this deceleration raises investor concerns, particularly regarding the impact of AI agents on the need for human software licenses.

      To tackle pricing challenges, Salesforce has introduced a consumption-based model for Agentforce instead of the traditional per-seat pricing, charging for "agentic work units." It has processed nearly 20 trillion tokens, converting them into over 2.4 billion units. The key question remains whether this model can grow sufficiently to compensate for the fundamental threat to seat-based revenue.

      Smaller clients exemplify both the potential and costs associated with this model. For instance, the city of Kyle, Texas, implemented Agentforce for its 311 service, managing over 12,000 resident requests since March 2025 with a nearly 90 percent first-call resolution rate. Bloomberg reported the city doubled its Salesforce expenditure to $300,000, which may be a reasonable investment for a fast-growing municipality. However, for larger enterprise customers considering similar decisions at scale, the economics remain uncertain.

      The competitive landscape is intensifying, with SAP launching its Autonomous Enterprise featuring over 200 AI agents in partnership with Anthropic at Sapphire 2026. Other companies like ServiceNow, Google, and Microsoft are also developing similar agent platforms. The question is no longer if AI agents will transform enterprise software, but whether Salesforce can uphold its position as the market alters around it.

      Benioff responded with notable confidence, setting a new revenue target of $60 billion by fiscal 2030. He has also announced $50 billion in share buybacks, indicating the company's belief that its stock is undervalued. The transformation of Slack into an agentic platform, including over 30 new AI capabilities and mandatory bundling with every new Salesforce account starting this summer, is part of this strategy.

      Nevertheless, this does not resolve the fundamental conflict. Salesforce is asking customers to invest