Health systems are racing to adopt artificial intelligence. But can they prove their value?
Health systems are racing to adopt artificial intelligence. But can they prove their value?
By Emily Olsen | Published: 2025-10-27 17:37:00 | Source: Healthcare Dive – Latest News
LAS VEGAS — Health systems are striving to apply artificial intelligence across their organizations, but how to measure the returns on those investments isn’t always clear, experts said at the HLTH 2025 conference last week.
AI technology, which includes ambient documentation clerks, summarization tools, and revenue cycle management products, costs money and time to put into practice. Experts say it’s not always clear how AI produces a dollar return on investment, which is important for justifying expenditures and measuring success.
“Using AI to simplify management, using AI to help productivity is absolutely there. There are still questions about the return on investment on it and getting to the bottom line,” said Tom Bales, consulting president of health and wellness services at PricewaterhouseCoopers.
At the same time, Health systems are also facing increasing financial pressures. This summer, President Donald Trump signed a massive tax and policy bill that will lead to historic cuts to Medicaid, likely culling millions of people from the safety net program.
On the other hand, more generous financial aid for people who buy coverage on the Affordable Care Act exchanges — an issue that was the focus of the weeks-long government shutdown — is set to expire at the end of the year, which would also push more people out of coverage.
More uninsured Americans means less revenue for providers and an increase in uncompensated care.
However, the healthcare industry is hoping that AI will shake things up on some of the biggest challenges it faces, including workforce shortages, provider burnout, and lengthy administrative tasks. Amid the challenging economic environment, healthcare companies are prioritizing AI investments that are more likely to improve profit margins and show clear returns, according to a survey published earlier this month by Klas Research and Bain & Company.
But because these returns are difficult to demonstrate, hospitals may have to get creative in measuring the value of AI, including by looking at time savings for providers, stronger employee retention, or improved patient satisfaction, experts say. These mitigating measures can then be linked to financial improvement over time.
“I think eventually we can track the majority of these things down to the bottom line, but you may have to connect more dots,” said Jennifer Goldsack, CEO of the Digital Medicine Society, an industry association.
Putting your ROI together
Measuring financial returns is always difficult, given the complex interplay of factors that complicate the final savings ratio to a given product. This also applies when it comes to AI, where the direct financial impact of a product depends on the tool, the costs needed to adopt it and the specific priorities of each provider, said experts at HLTH.
“In healthcare, especially on the healthcare delivery side, it’s really hard to apply a line of sight to what we might consider ROI,” Mickey Tripathi, Mayo Clinic’s chief AI implementation officer, said during a session Monday. “I don’t think any of us should really expect that you’re going to see cost cuts that a CFO can look at and say, ‘Oh, I saw it there.’
Measuring financial returns has become easier with some AI products. For example, revenue cycle management tools may allow health systems to see improvement in metrics such as time to collection, according to Sandra Johnson, senior vice president of client services at electronic health records provider CliniComp.
But the impact of other AI tools is less clear. There is limited evidence that artificial intelligence documentation assistants, which record doctors’ conversations with patients and create a clinical note, impact productivity and financial performance, according to a report published by the Peterson Health Technology Institute earlier this year.
“In healthcare, especially on the healthcare delivery side, it’s really hard to apply a line of sight to what we might consider ROI.”
Mickey Tripathi
Chief Artificial Intelligence Implementation Officer, Mayo Clinic
However, scribes likely reduce physician burnout, the report found — a major concern that can be exacerbated by the amount of time and effort needed to document care and do other work in electronic health records.
Experts say that if AI staff prevents a doctor from leaving the health system, it could be a significant cost saving. It can be expensive to replace a doctor Two to three times the annual salary of a doctorStudies indicate
There are other potential financial and efficiency gains from perimeter clerks, said Dr. Neely Jessel, chief medical officer at health technology company and electronic health records supplier Athenahealth.
For example, providers can complete notes more quickly and potentially see more patients. They may also be able to document services more accurately and bill for a higher level of services.
Patient satisfaction is another important metric for health systems. Patients tend to like the clerks around them in part because they… Doctors can be more present During appointments, instead of staring at the computer screen while taking notes. Documentation assistants may also help patients better understand their care plans, since doctors have to speak out loud during the exam to make sure the scribe captures the information, Jessel said.
“Providers are getting a lot better at, for example, (at) verbalizing the physical exam, talking about the plan, right?,” she said. “So it increases transparency and also makes for a more enjoyable experience for patients.”
How health systems are measuring the impact of AI
To successfully implement AI, health systems need to think critically about the biggest problems they want to solve and their key priorities, experts say. This plan should include how they will evaluate the AI tool, which can help them determine whether the project is successful – both financially and through other metrics.
“For me, Step Zero is very clear about what are we doing?” Why do we do that? What does success look like at this?” Mounir Odeh, chief data and artificial intelligence officer at Cedars-Sinai, said during a panel discussion on Sunday.
Cleveland Clinic evaluates AI using two types of metrics, according to Sonya O’Malley, managing director of business development and licensing at Cleveland Clinic Innovations, the marketing and innovation arm of the system.
This includes quantitative measures – such as reduced no-show rates or reduced documentation time, compared to a baseline measure – as well as more qualitative measures, such as feedback on the patient or physician experience.
Some tools will never be a source of income. But they may do other important things, like allowing a doctor to spend time on tasks that require his or her full training and experience, O’Malley said.
When an AI tool is rolled out to Ardent Health, success metrics must be clearly defined and measurable along with a baseline, according to Annika Gardenshire, the health system’s chief digital transformation officer.
Financial return on investment is almost always important, but it may be correlational rather than direct causation, she said.
If the tool only breaks even, the system must take into account other costs associated with it, such as implementation expenses or time spent by clinicians on the project when they could have been working elsewhere.
“If you get to a place where you’re just like, ‘Hey, this thing hasn’t even broken even.’ “She’s not even paying for herself,” Gardenshire said. Unless you’re doing something absolutely miraculous for patients, they’re gone.
ــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ
