Healthcare is entering a pivotal moment. Financial pressures1 are mounting across the industry, patients continue to struggle with a fragmented and expensive care experience, and organizations are searching for new ways to improve access,2 affordability, and outcomes. At the same time, advances in AI are creating opportunities to rethink how care is delivered, coordinated, and paid for. In the latest episode of the McKinsey on Healthcare podcast, McKinsey Senior Partner Patrick Finn, global leader of the Healthcare Practice, explores why this moment may represent more than another industry cycle—and what it could mean for the future of healthcare.
In a conversation with Querida Anderson, McKinsey senior editor for healthcare, Finn discusses the historically unusual convergence of financial pressure on both health insurers and clinical-care organizations and the forces reshaping the healthcare landscape. He explains why the current environment is creating both tension and opportunity and why industry leaders may need to rethink long-standing assumptions about care delivery and patient engagement.
The discussion also examines AI’s potential to move beyond improving administrative efficiencies and become a catalyst for broader care model redesign.3 While challenges remain, he argues that the industry’s willingness to embrace change may determine whether this moment becomes a temporary adjustment or a lasting transformation. The biggest question now is, who’s up for the challenge?
A lightly edited version of their conversation follows.
The simultaneous pressure on insurers and clinical-care organizations
Querida Anderson: The healthcare sector has been through many periods of upheaval over the course of its history. Most recently, of course, and fresh on everybody’s minds is the COVID-19 pandemic, which started in 2020. But as we sit here today in 2026, the situation feels different and unique in that both health insurers and clinical-care providers are under financial pressure at the same time. Historically, this hasn’t been the case—or at least not to this extent. What has changed in the system to create this simultaneous squeeze?
Patrick Finn: You’re right that the past decade or two have been a fairly unique time in the history of the industry. This squeeze has been driven by a changing funding environment and a changing demographic environment. If you look at the period after the Affordable Care Act, there was continuous injection of funds into the system to expand Medicaid, provide subsidies in the exchange markets, and provide enhanced subsidies during COVID-19 for those exchange markets or to directly support the delivery system.
Now, we’re experiencing the paring back of that growth in funding given other substantial budget priorities at a federal level. If you look in the employer markets,4 the runway for shifting costs to the employee is waning. So, in effect, there’s no more money, and that creates pressure on the system. We’re now seeing that across both payers and providers.
Querida Anderson: Do historical solutions still work? What are the new tactical options? Overall, what’s the strategy to ensure that patient care needs are met?
Patrick Finn: There are two answers to that question. First, it’s about improving the things that have been talked about for years, such as improving administrative efficiency and creating care models that get the right care to each patient at the right time and in the right setting.5 And so all the things that we as an industry have talked about for years—lowering administrative waste in the system, reducing unneeded care, getting patients into the right care settings, and improving health outcomes—all still hold. We all know those are real execution challenges. While there are great examples of places that have worked to improve each of those, the change and the scaling components remain difficult.
At the same time, we can’t have this discussion without talking about the incredibly disruptive technological revolution that we are living through. AI has the potential to accelerate the changes we know are possible but difficult to scale at any individual institution or for any individual patient. It also can help stakeholders restructure and rethink how care is funded and delivered. Rather than the traditional approach of thinking about how best to serve populations, the industry can move closer to the ultimate goal of care at an individual patient level: How does each patient get the right care at the right time in the right setting, consistently?
So the industry has to continue to execute on all the classic efficiency strategies that have been discussed for a long time. But now the question is, how does the industry do that better with AI?6 And how does it reimagine what’s possible with AI to deliver better experiences and outcomes that are more affordable for the patient?
Querida Anderson: Before we discuss AI on a deeper level, let me ask you about simultaneous financial pressure that health insurers and health providers are facing, as I imagine this is changing the payer–provider dynamic. From your point of view, what does that dynamic look like on the ground right now? Are we heading toward more conflict or new collaboration? How are you trying to steer toward collaboration and away from conflict?
Patrick Finn: You can read in newspapers about the conflict that is happening more and more on this front. In some ways, that is a natural outcome of both payers and providers being under pressure and seeking solutions—and the ongoing solutions may not be the same for both of them in terms of the classic rate negotiations that go back and forth or the programs and standards of care and access we are looking for.
In the short term, as everyone tries to figure out that optimal intersection of affordability, access, and quality, there is naturally going to be tension. As that tension yields outcomes, I think it will ultimately result in more collaboration. We need to view that tension as a natural part of the process in getting to a different, more efficient frontier that improves access, affordability, and quality in a way that is sustainable.
AI as a catalyst for innovation in care delivery and funding
Querida Anderson: Let’s talk more about AI. It’s obviously going to be a key tactical and strategic solution. The only question is, where is it truly going to move the needle and how quickly? So let’s start with the industry side of things and then explore benefits for patients. Where do you see the most potential for AI to make a difference?
Patrick Finn: For most of us in our daily experience across industries, we’re seeing that the potential of AI is amazing. And for healthcare, this technology, unlike many before it, is almost tailor-made to the challenges of our industry. The challenge has been how to use the massive amount of data that exist in healthcare and the massively complex interactions between those data points to really personalize a care pathway for an individual—because each of us is different. We’re different in our socioeconomic backgrounds and the geographies we are in and, therefore, in the care delivery that is available to us. We’re also different in our histories and our genetic makeup. AI is intrinsically capable of factoring these differences into its analysis. Lots of risks to work through, but the technology is intrinsically capable at its core.
There are the classic things you would expect to improve with AI, such as the responsiveness and cost of call centers or improving processes—particularly on the provider side in areas such as revenue cycle management7 or on the payer side in claims processing functions. We’re already seeing rapid adoption of AI for these areas. For example, in some places, more than half of physicians use ambient listening,8 in our experience. I think adoption is only going to accelerate.9
Where the potential for AI gets exciting is in how it could be used to fundamentally restructure the delivery of and payment for care. We’re starting to see progress in areas such as access and scheduling, which are now much more agentic-driven,10 consumer-friendly experiences to help patients get access to primary care or specialist care and make that process more seamless. There has also been more agentic-driven preappointment outreach to make sure patients are appropriately prepared, both clinically and from a documentation standpoint, so that time spent is more productive for the patient and the clinician.
In the clinical setting, AI has the potential to help clinicians have a more holistic view of each patient, for example, with relevant information on similar patients or the treatment options available to them. The clinician continues to be the decision-maker, but they have substantially more support.
And then in terms of prior authorization, payment processing, and disputes, AI can help clinics handle those in real time. With access to medical records, it can approve payments immediately and maybe even pay in real time, which allows better navigation for the patient, better economics for the provider, and less administrative costs for the payer. It also helps guide patients to the right care at the right time in the right location, as opposed to what can be a painful back and forth over days and weeks as patients and physician’s offices work out what the next steps are.
Rising to the moment for care model redesign
Querida Anderson: You’ve already touched on much of how AI can help patients. So beyond AI, how can other technologies or redesigned care models improve patient experiences and outcomes?
Patrick Finn: The interesting question is, who will take the lead on holistic care model redesign? This could include health and wellness prior to the need for care, other types of preventative care, routine care, or specialist care. Every industry participant has a role to play in redesigning care models and the change management challenge that lies ahead to ensure these tools are used as companions along that care journey.
There’s an argument that the payers could take a lead11 in that, and there’s an argument that large health systems could take a lead in that. The interesting thing to watch is whether this technology will be an enabler for new entrants to disrupt how care is delivered in the United States. We can point to a decade plus of many digital natives and others attempting to disrupt the healthcare system—and a variety of reasons for why that hasn’t taken off. An important reason is that managing the complexity of the healthcare delivery and payment system is a real challenge.
So can this technology help alleviate these challenges? And will it enable new entrants to be companions to patients navigating their care journeys? We are already seeing some of that informally. Patients Googling their symptoms and possible diagnoses is evolving into patients using AI platforms to ask questions about their health. So the next question is, will we have new entrants create patient companions that use AI to help guide patients to where best to get diagnosed, where best to seek care, how best to fund that care, et cetera?
Querida Anderson: Zooming out, do you think this is a moment for a temporary margin reset or real structural change in the US healthcare system?
Patrick Finn: If you had asked me that question 18 months ago, in the absence of any regulatory change, which is historically what has moved the industry, I would have said, “This will be a temporary reset. It will get worked through. There will continue to be ongoing shuffling and consolidation and new start-ups but not a fundamental change.”
However, the speed at which we are seeing the capabilities of AI advancing ought to cause us to question whether this is just temporary or a bigger change for the industry. The question is, which industry participants are going to be first to move to disrupt themselves and think about different delivery models, different funding models, and different patient or member engagement models?
The potential is there, so it’s a question to industry participants of who’s going to take that leap and opportunity to really restructure the way care is delivered and how members and patients are supported.
