Agentforce's Marketing-Reality Gap: The Bloomberg Investigation's Named Customers and Salesforce's Response

Agentforce's Marketing-Reality Gap: The Bloomberg Investigation's Named Customers and Salesforce's Response

The Bloomberg investigation behind the "Agentforce ROI gap" narrative is now pinned down: "Salesforce Touts AI Promise Over Reality in SaaSpocalypse Fight" by Brody Ford, published May 22, 2026 (read this cycle via the Mercury News syndication; Bloomberg's own page is paywalled). Its method: named customers, interviews, and documents "seen by Bloomberg." Three showcase customers anchor the story:

• University of Chicago Medicine — an October 2025 promo video showed patients refilling prescriptions, booking appointments, and getting parking tips via Agentforce. But "the scenes are largely aspirational — little of that AI functionality is live¹. Patients who call the hospital system today are greeted with keypad-selection menus and routed to human schedulers." Delays were attributed to product glitches and internal compliance sign-offs; the health system "pays Salesforce millions of dollars per year" for multiple products.
• Williams-Sonoma — a December user-conference demo showed an Agentforce-powered support phone line ("Williams-Sonoma's most popular channel is voice"). But "about half a year after that presentation, Williams-Sonoma's phone line isn't actually using Agentforce. Even its online chatbot is unable to place or modify orders without a human customer support agent." CTO Sameer Hassan said the demos reflect "possibilities," with a phone-system go-live targeted for the holiday season.
• Finnair — a video showed Agentforce automatically rebooking a flight; the airline's website says such capabilities are only "planned for future development," blocked by a separate system that is "missing agentic capabilities."

Real production wins in the same piece: SharkNinja cut service calls 20% this year with automated troubleshooting; Adecco runs "tens of thousands of initial written screener interviews" through Agentforce; Kyle, Texas more than doubled Salesforce spend to $300,000 running resident-complaint intake (potholes, graffiti); PepsiCo boosted Salesforce spend 10% (chief strategy officer Athina Kanioura).

Salesforce's response: Benioff — "Whatever technology we have ever marketed, we have always delivered," and "I don't want any customer to ever call me and say — 'hey, I thought I was buying this, but it's not available.'" Agentforce EVP Madhav Thattai: "You're going to see customers start with simpler use cases… they kind of realize this full autonomous vision." Operationally, Salesforce is tying salespeople's compensation to actual customer usage, offering free or heavily discounted deployment help, and leaning on the $8B Informatica acquisition for the data layer. Analysts: Gil Luria (DA Davidson) — "many companies have gotten to promising a lot more than they actually have… they've just been so loud about it"; Valoir's Rebecca Wettemann on the fear that "the AI is going to run off in the middle of the night and refund a bunch of transactions or run up a huge bill."

Re-surfaced at Dreamforce 2026: Tech Times (Sep 17) reported that "a separate Bloomberg investigation published before Dreamforce found enterprise customers reporting that Agentforce… is producing outcomes that materially miss what Salesforce's marketing promised," alongside "a Goldman Sachs analysis [noting] that agent token economics are shifting rapidly, creating unstable ROI calculations across different deployment contexts" — and that Koa, the new CRM reasoning model, launched on Salesforce's own benchmark with no independent validation.

Context for the trajectory: at publication, Agentforce had 29,000 deals and $800M ARR; the stock was down 21% in 2025 and another 30% in 2026 amid the "SaaSpocalypse" (TNW counts ~$285B of SaaS market cap lost in a single 48-hour February window). The piece cites Apple's $250M May 2026 settlement over exaggerated Apple Intelligence claims as the emerging legal benchmark for AI marketing gaps. ARR has since crossed $1.2B (Q1 FY27) — tracked in Agentic AI Market Size and Growth: Salesforce's Agentforce Metrics from Q2 FY27 Earnings to Dreamforce 2026 and the $2B ARR watch.

What it means: the ROI-gap debate now has a named-customer evidentiary record rather than survey sentiment. The pattern — information-collection agents live, action-taking agents stalled on compliance and integration — matches the bounded-autonomy production evidence in Enterprise Case Studies: What 150 Evidence-Screened Agent Deployments Actually Show (September 2026) and the adoption/value divergence in The Enterprise AI Agent Production Gap: Adoption Is Near-Universal, Measurable Value Stays Rare (Late-September 2026 Wave). It also hands Salesforce's rivals (SAP's 200-agent Autonomous Enterprise, ServiceNow, Microsoft) a credibility wedge.