How AI Can Assist with Workload Balance for Nurses

AI should be implemented thoughtfully and built in a culture of trust, says this nurse leader.

In an era where efficiency and output are the priorities in healthcare, it can be easy to overlook the burdens of care on the nurses who deliver it.

In an environment where workload imbalances can lead to burnout, mistakes, and worse patient outcomes, it's important for CNOs to do what they can to help maintain equilibrium. According to a time motion study from 2018, nurses spend approximately 35% of their time in the patient room and about 25% of their time on documentation, though other studies report anywhere from 26.2% to 41%.

According to Marisa Quinn, director of nursing for infusion services at UCSF Health, the value of AI powered tools in staffing and workload balance is primarily that it reduces administrative burden for nurses and adds objectivity and visibility to complex real-time decision making.

"I would say it only works as a copilot and not as a replacement for clinical judgement," Quinn said.

AI for balancing workload

At UCSF Health, the organization implemented an AI-powered tool to support nurses and give them a bird's eye view of their units on one screen, Quinn explained. Nurses can see patient assignments by nurse, and get a better understanding of workload balance across the units.

"The tool uses predictive analytics to suggest the best nurse for the next patient as they arrive at our center within parameters and guardrails that we've built," Quinn said.

Leadership implemented the tool at the request of the bedside nurses who felt that their assignments were unbalanced, Quinn explained.

"They didn't have visibility into what was happening on the units with all of their team members," Quinn said, "and the workload intensity, or how hard it was to take care of their patients, of each nurse's assignment was not being captured."

According to Quinn, the nurses don't get as overwhelmed or feel as rushed as they did before they started using the tool. The tool can signal potential bottlenecks and allows nurses to be more proactive with their schedules.

"I think that's what working smarter actually looks like for nurses," Quinn said.

Because AI cannot capture everything a human can, UCSF uses the tool with a 'human in the loop' approach. Quinn explained that the nurses and charge nurses using the tool are the final decision makers, and they have the option to reject or accept every suggestion that the tool gives them.

"We were very intentional about using it as a decision support tool and not as a decision maker for them," Quinn said.

Improving outcomes

After implementing the tool in 2024, the health system saw meaningful impact on workload balance, pacing of assignments, and operational visibility for charge nurses, Quinn explained.

"After three months of our initial pilot, we did see 9% improvement in workload balance, 8% improvement in nursing productivity, and three-fourths of the nurses said they had better pacing of their assignments," Quinn said.

Following those results, the organization spread the adoption across half of its centers in 2025. Quinn explained that the nurses are getting real time recommendations, but they still retain full control and final say over their patient and staffing decisions.

As for the financial impact, Quinn emphasized that the goal of this initiative was not cost savings, it was sustainability and the ability to grow and respond to demand with a workforce that was able to handle it.

"This ties [into] retention, engagement, [and] wellness," Quinn said. "That's where you're going to see the cost savings."

Overcoming obstacles

According to Quinn, the biggest hurdle in implementation was building trust with the nurses. The pilot started on a small unit that verbalized the workload imbalance issue.

"They needed to see that their clinical judgement mattered and that the technology was there to support them and assist, but not automate their workflows," Quinn said.

Incorporating any change to nurse workflows is difficult, but when technology gets involved, it can be even more worrisome for nurses. Quinn emphasized that the point is not to automate nurse workflows.

"It was just [about] building that trust and that eventually came from the transparency that the tool provided, and adoption followed from there," Quinn said.

For CNOs who might want to implement a similar strategy, Quinn recommended starting by listening to the problems that teams are sharing, determining if fixing those problems can move the needle on outcomes, and then getting technology involved when necessary.

"Nurses are here to take care of patients," Quinn said, "but having that shared expectation and bringing the process to the surface for all the teams to understand helped with confidence in the tool and [with] feeling like the [nurses] were supported."

Quinn also recommended engaging frontline staff as early as possible, since those leaders are going to be the ones rolling out these changes to their teams. CNOs should also be as transparent as possible. AT UCSF, leadership was able to develop the tool with the vendor based on feedback from frontline staff, which provided credibility, trust, and improved adoption once the tool was implemented.

"I think AI can be really powerful, but it has to be embedded in thoughtful process redesign and then built in a culture of trust," Quinn said, "and using that framework of aligning people, process, and technology so you can create sustainable change."