Nurses at HCA Healthcare hospitals say an artificial-intelligence scheduling system is producing exhausting assignments, uneven teams, and new administrative burdens that may affect patient care. Their accounts appear in an October 2 investigation by WIRED into Timpani, a tool HCA co-developed with Palantir.
HCA has deployed Timpani at roughly 130 of its 190 facilities since 2023, according to the report. The system uses forecasts of patient volume and staffing needs to generate nursing schedules. HCA has presented the software as a way to save managers time, reduce dependence on contract nurses, improve retention, and remove favoritism from shift assignments.
Several frontline nurses described a different experience. Florida critical-care nurse Amber Retzloff told WIRED that she requested 50 specific 12-hour shifts over four months but was assigned elsewhere more than half the time. She said the resulting patterns sometimes placed her on three consecutive workdays and left her mentally depleted.
The more serious allegations concern staffing quality, not personal convenience. Nurses interviewed by WIRED said Timpani can schedule too few people or create teams without enough experienced staff, especially on Sundays. Retzloff recalled working as the only senior nurse alongside four junior colleagues and said she delayed care for the sickest patients while helping the less experienced staff manage other cases. These are workers’ accounts; the report does not establish that Timpani directly caused patient harm.
HCA disputes the suggestion that its software is intended to reduce staffing at the expense of care. Company spokesperson Harlow Sumerford told WIRED that nursing leaders, rather than Timpani, make the final decisions. He said HCA’s aim is to match patient needs with caregivers who have the right skills while also considering employee preferences, and that the company is continuing to improve the tool using nurses’ feedback.
HCA also supplied figures supporting its position. The company said Timpani assigns nurses to about 1 percent of the days they specifically request off. An HCA innovation executive wrote in July that the system had sharply reduced the time managers spend building schedules, lowered reliance on contract labor, increased retention, and produced schedules that included a mix of skill and experience more than 98 percent of the time.
Nurses told WIRED that even infrequent exceptions can carry a large human cost. Missouri nurse Lee Barker said being assigned on a requested day off was previously almost unheard of at his facility and that he postponed a long-planned medical appointment after he could not trade one such shift. Nurses also said they now spend more time appealing assignments or searching for swaps, with disputed schedules reviewed by a centralized team rather than adjusted through direct conversations with local managers.
The investigation found concerns beyond HCA. Internal documents reviewed by WIRED showed complaints at Rayus Radiology after the imaging chain introduced a scheduling tool built on Palantir Foundry. In some cases, according to the report, the system connected requests to the wrong patient profiles, listed scans that differed from physicians’ orders, or generated inaccurate personal details. Rayus said those descriptions did not fairly represent its safeguards and emphasized that technology supports rather than replaces human judgment.
A Palantir spokesperson declined to be named in the WIRED report but said the company’s software presents information more usefully while customers remain responsible for their data, policies, and decisions. That division of responsibility is central to the dispute: hospital leaders describe an aid whose outputs remain subject to human review, while nurses describe a centralized system whose recommendations are difficult to challenge in practice.
The accounts do not prove that algorithmic scheduling is inherently unsafe, and WIRED noted research suggesting such systems can be useful. They do show that performance claims based on efficiency or aggregate staffing measures may miss the operational details clinicians experience on individual shifts. In a hospital, whether a schedule works depends not only on filling positions, but also on fatigue, experience, continuity, and whether staff can correct a bad assignment before it affects care.