HCA nurses say Palantir-built scheduling AI overrides preferences and thins skill mix

WIRED reports that nurses across HCA Healthcare, the largest US hospital chain, describe Timpani, an AI scheduling system co-developed with Palantir and rolled out since 2023 to about 130 of roughly 190 locations, as routinely ignoring preferences, stacking exhausting consecutive shifts and under-weighting experienced staff, especially on Sundays. Timpani runs on Palantir Foundry and forecasts patient volume and staffing.

Critical care nurse and National Nurses United leader Amber Retzloff says more than half of 50 requested 12-hour shifts were overridden, and describes delaying care for the sickest patients to coach newer nurses because she was the only senior nurse on shift. Nurses say days they reserved off are now scheduled and appeals go to headquarters in Nashville without discussion. A Missouri psychiatric unit grievance is pending. National Nurses United represents about 10,000 of HCA’s roughly 100,000 nurses and says its information requests about the technology were rejected.

HCA spokesperson Harlow Sumerford says nursing leaders, not Timpani, make final decisions, and denies the tool is designed to cut staffing at patients’ expense. Innovation executive Michael Schlosser has claimed large reductions in manager scheduling hours, less reliance on contract nurses, higher retention and that over 98% of schedules mix skills and experience. Palantir’s position is that it presents information and customers own data and decisions.

A former HCA data science manager, Angelique Russell, filed a lawsuit in July alleging data deletion that blocked audits; HCA had not formally responded in court when WIRED published. Separately, staff at Rayus Radiology reported Palantir Foundry scheduling errors that tied orders to the wrong patients or invented details, a different failure class involving extraction from messy inputs. Similar worries about workers absorbing AI errors surfaced in hospitality the same week (customers citing chatbots).

Efficiency gains for managers and labor costs can coexist with worse bedside conditions. When coverage is measured as headcount rather than a map of competence, skill-mix failures follow. “Leaders make the final call” offers limited protection if an automated draft becomes the hard-to-contest default and appeals are slow and distant.

Hospital AI works best with auditable objectives nurses help set, such as honoring reserved days, enforcing seniority mix and showing confidence, plus logs that cannot be casually deleted and local authority to fix bad drafts quickly. Palantir’s health push coincides with strong sales and separate controversy over its UK NHS work. The arbitration outcome and the Russell suit will test the claims. More in our healthcare coverage, future of work coverage and enterprise coverage.