For years, artificial intelligence (AI) in healthcare has been pitched to employers as a cost-saving tool that could be used for faster diagnoses, smarter triage and less paperwork. However, recent data reveals a more complicated story. According to PwC’s annual medical cost trend survey, health plans are now projecting that commercial healthcare costs will rise 9% in 2027, with providers’ adoption of AI-enabled revenue optimization tools listed among the leading inflators.
This article explores how AI and healthcare have become increasingly intertwined and examines how this shift has made AI a leading driver of rising healthcare costs. It also outlines what employers can do to budget, negotiate and evaluate AI-related vendors more effectively going forward.
History of AI and Healthcare
AI’s footprint in care delivery has grown quickly, not just through the clinical tools that usually get the spotlight, like diagnostic algorithms, imaging analysis or AI-powered virtual care. Those applications tend to dominate the conversation because they are visible, dramatic and easy to imagine improving outcomes. But the use case actually driving costs right now is far less glamorous since it is administrative.
AI scribes and coding assistants that listen to or read clinical visits and automatically draft notes, suggest diagnosis codes, and flag billable complexity have quietly become one of the most consequential applications of AI in healthcare. The scale of investment behind this administrative shift is significant: UnitedHealth Group, the country’s largest health insurer, is putting $1.5 billion into AI this year alone, with company leadership describing it as a reimagining of how the organization works, not simply a productivity boost. Much of that spending is aimed squarely at administrative functions, such as prior authorization processing, pharmacy benefit approvals and claims handling, rather than the clinical algorithms that tend to get more public attention.
These tools have spread fast because they solve real problems. Clinicians spend a huge amount of time on documentation, and AI can do much of it faster and more thoroughly than a person can. That thoroughness, though, is exactly what makes these tools financially significant. When AI captures every detail and condition a clinician mentions during a visit, providers can legitimately code that visit as more complex and bill more for it, even if the treatment itself hasn’t changed.
How AI Became a Top Cost Driver
According to PwC’s survey, nearly 70% of health plans rank AI-driven documentation and coding tools among their top three cost inflators for next year, and about 1 in 5 call AI the single biggest driver of all. The core issue isn’t more people getting more care; it’s billing accuracy. Costs are rising mainly because claims are being coded as more complex and paid at higher rates per claim, not because the number of services delivered is spiking. The following trends are reinforcing this pattern:
- Providers are under financial pressure. Hospitals are dealing with rising costs and cuts to public health programs, so they have a strong incentive to make sure they’re capturing every dollar they’re legitimately owed. AI makes it much easier to do that consistently, across every visit.
- The payment system often rewards volume and complexity, not results. Providers get paid based on how many services they deliver and how complex those services are coded, not on how healthy their patients turn out to be. Since AI is good at maximizing both volume and coding complexity, it naturally drives spending up, even when the actual care being delivered hasn’t changed.
- Billing rules weren’t built for AI. Most payment systems were designed around what a human could realistically do in a set amount of time. AI doesn’t work like that. It’s cheaper to scale and can handle tasks that no billing code ever accounted for. Until payment policy catches up, this gap between how AI performs and how billing was designed will likely keep pushing costs higher, without any clear improvement in patient outcomes.
There are reasons to expect costs to improve over time. Administrative work makes up a large share of total health spending, so automating it could eventually cut overhead and reduce provider burnout. Those savings could eventually translate into lower costs overall. Many experts also expect the current focus on AI for billing and paperwork to fade as the technology matures, with more value shifting toward AI that catches health problems early and keeps people healthier, since that’s ultimately cheaper than treating serious illness later.
The catch is that these improvements tend to make the entire insurance industry more efficient, rather than giving any one health plan a lasting edge over its competitors. In other words, this is more likely to play out as a slow, industrywide shift than a quick fix any single employer can count on seeing next year.
What Employers Can Do
For plan sponsors, AI-driven coding intensity should be treated as a real and growing factor in next year’s budget conversations, alongside familiar drivers like hospital labor costs, prescription drug spending and behavioral health utilization. A few practical steps can help:
- Push for visibility into your own claims data. Ask your health plan or third-party administrator how coding patterns, not just headline trend numbers, are shifting in your specific population.
- Ask about the payment model direction. Find out whether your carriers and provider networks are moving toward models that reward outcomes rather than volume, since that shift directly affects how AI-driven documentation gets used.
- Evaluate vendor claims on the evidence. Don’t assume every AI-branded health solution automatically saves money. Vet these tools on their own track record.
Takeaway
AI is reshaping how healthcare is delivered, but in its most widely adopted form today, documentation and billing assistance, early evidence shows it’s adding to costs rather than reducing them. The outlook that AI will lower healthcare costs through earlier, more efficient and more personalized care is real. But most experts agree it will take time, as well as possible changes to how care gets paid for, before that outlook outweighs its current price tag.
Contact us today for more employee benefits resources.
This article is not intended to be exhaustive nor should any discussion or opinions be construed as professional advice. © 2026 Zywave, Inc. All rights reserved.
