This insight was featured in the February 20th, 2025 edition of the AI Catalyst Pulse.
First in a four-part series examining the critical strategic dilemmas healthcare executives must navigate in 2025.
Every dollar your health system invests in AI has to serve one of two goals: securing a larger share of today's healthcare dollars, or transforming how care gets delivered over time. This is healthcare's first fundamental AI dilemma in 2025: Do you prioritize shorter-term value capture or longer-term value creation?
(A small disclaimer: All of the dilemmas we’ll explore in this series are, to some extent, false binaries; you can do some value capture while also pursuing value creation. That said, we’ve heard repeatedly from AI Catalyst members that budgets are tight and that they must make hard trade-offs. Our goal is to provide frameworks that help you weigh the pros and cons of each approach and give clear guidance to your team.)
The case for value capture: You need immediate wins in the ‘AI arms race’ with your payers
The revenue cycle arms race is accelerating. One survey found that final denial rates for inpatient care surged 51% between 2021 and 2023, and denials have only increased since then as payers have deployed increasingly sophisticated AI tools. It’s debatable whether payers’ tools are terribly accurate — in one recent case, UnitedHealth was accused of using an AI model with a 90% error rate to deny Medicare Advantage claims — but their impact is undeniable.
Fortunately, you're not helpless in this fight. Health systems are deploying their own AI tools and seeing dramatic results:
One AI Catalyst member boosted revenue by $15 million in year one using AI-powered CDI.
Emory Healthcare cut appeals letter writing time by 90%, with 75% of AI-drafted appeals succeeding.
Mass General Brigham's AI tools identified 90% more patients with adverse social determinants of health, potentially improving the likelihood of prior authorizations.
And you can achieve at least some of these gains fast. Some organizations report that they’ve been able to begin drafting AI-powered appeals letters within hours of launching a vendor tool.
But there's a catch: Because your payers are also deploying AI to advance their own bottom line, your early gains may not last. The health system that saw a $15 million boost in first-year CDI revenue, for instance, saw the gains drop to $7 million in year two as payers adapted.
As Dr. Robert Wachter of UCSF pithily framed the challenge to the New York Times, “[Payers’] AI will deny our AI, and we'll go back and forth.”
The case for value creation: You could transform healthcare – and see compounding gains over time
Some health systems are making a longer-term bet: focusing their discretionary AI investments on transforming care delivery. In other words, rather than engaging in a zero-sum battle with payers over every dollar currently spent on healthcare, they’re trying to use AI to provide better, more efficient, or just plain new care.
The early results are compelling. For instance, physicians at Texas Health Resources who use DAX CoPilot to automate their clinical documentation are saving more than five hours weekly — time they're using to see more patients while actually increasing face-to-face interaction. At WellSpan Health, AI virtual assistants are handling routine patient outreach and education around colonoscopies that once consumed valuable nursing hours.
Unlike revenue cycle gains, which evaporate as payers adapt, these operational improvements compound over time. Staff get better at working with AI tools. Patients experience better care. The benefits multiply.
Which should you choose: Value capture vs. value creation?
While you’ll undoubtedly need to do both some value capture and some value creation, trying to do both aggressively risks muddling your message and confusing your staff. You need a clear strategic emphasis.
Our guidance: Consider prioritizing value capture if you're under immediate financial pressure and need rapid ROI, or if you’ve seen sharp recent increases in your denial rates (which might suggest that your payers are already beating you in the AI “arms race”). In that case:
Identify the “pain points” in your revenue cycle (for example, prior authorizations or CDI) and choose AI partners who target those specific areas.
If in doubt, start with appeals letter automation, which typically shows the fastest returns.
Use today’s “value capture” efforts to build the AI governance and technology infrastructure that will be required for tomorrow’s “value creation” efforts.
On the other hand, consider prioritizing value creation if you have short-term financial breathing room and your payer mix (for instance, a high proportion of traditional Medicare) makes you less vulnerable to AI-driven denials. This path also makes sense if your organization is already excited about AI's potential, since you'll need broad buy-in for transformational efforts. In that case:
Focus first on proven use cases like clinical documentation and patient flow optimization, where other systems have already demonstrated lasting impact.
Build AI literacy across your organization deliberately, as large-scale transformation requires staff who understand both the potential and limitations of AI tools.
As you demonstrate success with initial projects, measure both financial and operational impacts. You'll need this data to justify continued investment in transformation over short-term revenue gains.
Finally, here’s an intriguing data point: When we posed the “value creation vs. value capture” dilemma to health system CEOs and board chairs at The Health Management Academy's Trustee Summit, 82% favored value creation. That's a contrast to what we’ve seen from CIOs and COOs, who are more evenly split between the approaches.
The gap suggests that while healthcare’s top leaders envision AI primarily as a transformative force, the immediate pressures of revenue capture are harder to ignore on the ground.
Questions to consider:
Which specific revenue cycle processes are losing you the most money to denials? That's where you need AI defense first.
If you choose value creation, how will you explain to your board why you're accepting short-term revenue losses for long-term transformation?
How will you create organization-wide understanding of your strategic choice and its implications?