Salesforce

Salesforce is facing growing scrutiny after a recent Bloomberg investigation raised questions about the gap between Agentforce marketing and real-world deployment. The report focused on Salesforce’s flagship “agentic AI” platform, Agentforce, and highlighted cases where promotional demos appeared far ahead of what customers are actually using today. One example cited was UChicago Medicine, featured in a 2025 Salesforce video showing patients seamlessly using AI for prescription refills, appointment scheduling, and parking assistance. According to Bloomberg: • Many of those advanced capabilities are still being rolled out in phases or remain in testing • Patients still primarily interact with traditional phone menus and human schedulers • Some chatbot functionality is not yet broadly visible in production To be clear: this does NOT mean Agentforce is fake. Salesforce has reported massive growth: • Agentforce ARR reportedly reached \~$800M by Q4 FY2026 • Combined Agentforce + Data Cloud ARR exceeded $2.9B • The company says it has closed tens of thousands of AI-related deals The bigger issue is one the entire AI industry is now facing: AI demos are advancing faster than enterprise deployment reality. In highly regulated industries like healthcare, deploying autonomous AI systems at scale requires: • compliance reviews • data governance • integrations with legacy systems • human oversight • phased rollout strategies That creates a widening gap between: what AI vendors market today what customers can safely operationalize today This isn’t unique to Salesforce. Across enterprise software, many “AI agent” products still require heavy customization, structured data, workflow tuning, and human escalation layers before they deliver fully autonomous outcomes. The Bloomberg piece lands just days before Salesforce earnings, where investors will likely focus heavily on: • actual Agentforce adoption • production usage vs pilot deployments • monetization • customer ROI • AI revenue durability The broader market debate is becoming increasingly clear: Are we seeing true enterprise AI transformation… or a temporary hype cycle where expectations are outrunning implementation reality?