AI was supposed to make workplace scheduling easier. For some nurses in the US, however, it appears to be creating another problem they have to clean up. Nurses at HCA Healthcare, the largest hospital chain in the United States, say an AI-powered scheduling system called Timpani has made it harder to manage their shifts, personal lives and workloads.
The software was developed by HCA in collaboration with Palantir and has been rolled out at roughly 130 of HCA's 190 hospitals since 2023, according to WIRED.
Nurses Say AI Is Ignoring Their Scheduling Preferences
Amber Retzloff, a critical care nurse in Florida, told WIRED that she requested 50 specific 12-hour shifts over a four-month period. According to her account, Timpani assigned her to different shifts more than half the time.
'No one is happy with Timpani,' Retzloff said. 'Everyone complains all the time.'
Retzloff added that the system frequently placed her on consecutive shifts, including back-to-back-to-back working days, leaving her mentally drained and 'exhausted.' She said she did not expect every preferred shift to be granted, but believed the software performed worse than the manual scheduling system previously used by managers. Other healthcare workers have raised similar concerns about the system.
HCA is 'letting AI dictate,' Retzloff said. 'We are overriding clinical judgment and the human-to-human piece of health care with an app.'
Workers represented by 1199SEIU in Florida have said Timpani has denied paid time off, failed to honour preferences and changed schedules in ways that conflicted with workers' personal responsibilities and appointments. The union also said some employees were removed from schedules, resulting in lost wages.
Why Are Nurses Angry About Timpani?
The dispute is not simply about getting the wrong day off. Nurses say scheduling software cannot always understand the realities of working on a hospital ward, particularly when patient needs change rapidly. One concern raised by nurses is the balance of experience on a shift.
WIRED reported that Retzloff had been scheduled as the only senior nurse on some shifts, alongside less-experienced colleagues. She said this could leave her spending time supervising junior staff rather than concentrating on the patients who needed the most attention.
That has turned the debate over AI scheduling into a much bigger question: can an algorithm understand the human factors involved in deciding how many nurses a hospital needs at a particular moment?
National Nurses United, the largest US nurses' union, has argued that automated staffing systems can move important decisions away from frontline clinical judgement. The union has organised protests against Palantir's involvement in healthcare and called for greater transparency around the technology.
HCA Says Nurses, Not AI, Make the Final Decisions
HCA has pushed back against the criticism. The hospital company says Timpani is designed to help nursing leaders create schedules by considering factors such as employee preferences, skills and staffing requirements.
HCA maintains that nurse leaders can review and edit the schedules generated by the system and remain responsible for final decisions.
A spokesperson previously told healthcare industry media that the technology is intended to reduce administrative work and help nursing leaders accommodate employees' scheduling preferences. HCA has also said Timpani does not make the final scheduling decisions and does not use patient data.
That creates an unusual problem at the heart of the controversy. If a schedule is generated by software but a human manager ultimately approves it, who is responsible when the schedule goes badly wrong?
AI Scheduling Was Meant to Solve a Human Problem
Hospitals have long struggled with staffing shortages, complex rotas and the administrative burden of scheduling nurses across different shifts. That makes healthcare an obvious target for automation.
In theory, software can process thousands of variables far faster than a human manager. It can account for availability, qualifications, requested shifts and staffing requirements while generating a rota in a fraction of the time.
But nurses argue that hospital staffing is not simply a mathematical puzzle. A ward can have the required number of people on paper while still lacking the right combination of experience for the patients being treated.
A critically ill patient may require an experienced nurse. A junior member of staff may need supervision. Patient numbers can also change during a shift. Those factors are difficult to capture in a scheduling system.
Nurses Have Taken Their Concerns to the Streets
The dispute has escalated beyond individual complaints. National Nurses United organised demonstrations in eight US cities in August as part of a campaign against Palantir's growing involvement in healthcare technology. The union said its concerns included automated staffing systems, data practices and the potential replacement of clinical judgement with software.
The demonstrations followed months of criticism from nurses and healthcare workers. The union has argued that healthcare systems should prioritise hiring enough nurses and improving working conditions rather than relying on technology to solve staffing problems.
Palantir, meanwhile, has become increasingly involved in healthcare technology in both the US and UK.
In Britain, more than 44,000 people had filed legal objections by 30 September to NHS England's use of the Palantir-operated Federated Data Platform, according to The Guardian. The objections concern how personal health data is handled, although NHS England says healthcare organisations remain in control of their data and suppliers cannot use it for their own purposes.
The controversy surrounding Timpani is therefore part of a much broader debate about how far AI should be allowed to influence decisions in healthcare.
The Bigger AI Workplace Problem
The Timpani controversy highlights an uncomfortable reality about workplace AI. Automation does not necessarily remove human work. Sometimes it simply changes the work.
Instead of spending hours creating schedules, managers and employees may end up spending that time checking, correcting and challenging decisions made by software.
For nurses, the stakes are particularly high because a scheduling error does not just mean someone has an inconvenient Tuesday.
It can mean a nurse is exhausted after consecutive long shifts, a ward has an imbalance of experience, or a worker has to rearrange their life because the system failed to honour an agreed preference.
As AI moves deeper into workplaces, the question may no longer be whether an algorithm can make a decision faster than a human. It may be whether the people left to fix the algorithm's mistakes are actually better off.
