The Trump administration launched WISeR (Wasteful and Inappropriate Service Reduction) in January 2026, an AI‑driven prior‑authorization pilot for Medicare services in six states (New Jersey, Ohio, Oklahoma, Texas, Arizona, Washington). The system contracts private vendors to apply machine‑learning models that decide whether procedures such as nerve stimulation, epidural steroid injections, cervical fusions, skin substitutes, incontinence and impotence treatments receive approval. According to documents obtained by the Electronic Frontier Foundation, Innovaccer initially auto‑approved all requests to avoid backlogs, while Zyter struggled with Medicare Part A vs. Part B data mismatches. Virtix reviewed 6,096 requests in a March 30 weekly report, approving 2,863 (47 %) and denying 3,233 (53 %), far exceeding the target 72‑hour decision window; some cases lingered over 80 days, prompting a CMS Corrective Action Plan that Virtix says reduced its average turnaround to 1.18 days. CMS reimburses vendors 25 % of the estimated cost of each denied service, with a modest quality‑score penalty that pays 90‑95 % of that share even when denials are inappropriate. The program, which faced provider complaints of patients crying in pain due to delayed kyphoplasty/vertebral augmentation, is slated to expand to oncology, air‑ambulance transport, advanced imaging, cardiac devices and molecular labs through 2031.