Ninja radar.

Commonwealth Bank of Australia cut more than 40 customer service roles and replaced them with an AI voice bot, then reversed the cuts after the system “was unable to cope, which led to an increase in calls” (CNBC, 2026-07-01, citing Australia’s finance sector union and an ABC report). The union called the rescission “a massive win.” The underlying cuts date to 2025; the fresh reporting is CNBC’s July 2026 retrospective on AI-layoff reversals. Single independent source in this flash; the union statement and ABC report are the primary trail for a deeper case file.

IBM automated its human resources functions with AI that “handled around 94% of routine requests but was unable to meet the other 6%, which included ethical dilemmas” (CNBC, 2026-07-01). The company then announced plans to triple its US entry-level hiring across all business units in 2026. Its chief human resources officer, Nickle LaMoreaux, framed the rehiring as a pipeline problem: “If we don’t continue to invest in entry-level hires, what happens in three-five years? There’s no pipeline.” The 94 percent figure is company-originated and relayed by CNBC; single independent source in this flash.

Salesforce announced Agentforce Help Agent with pay-per-resolution pricing: customers are charged only when the agent resolves an issue, generally available July 2026. The proof point is Salesforce’s own deployment: the company says Agentforce on help.salesforce.com “handled 4.3 million inquiries and resolved 70 percent of them” (announcement dated 2026-06-25), a figure CIO relayed as “resolved 70% of them” the same day. Both figures originate with Salesforce and are not independently audited; the deployer and the vendor are the same company, so read this as a vendor’s self-deployment claim with an unusually concrete pricing commitment attached. The pay-per-resolution model itself is the newsier fact: it prices the product on outcomes, which invites external measurement.