AI in the revenue cycle isn’t at a tipping point; it’s a sticking point.
Revenue cycle leaders at health systems across the country are being given AI mandates. At Sift, we’ve heard from multiple revenue cycle leaders that their CEOs and even CFOs are asking, “What’s our AI strategy?” or issuing directives like “60% of our initiatives must include AI this year.”
Revenue cycle leaders have taken these mandates seriously. They take vendor calls. They pull up McKinsey papers. Some of them take second and third calls. They build a shortlist and they start to get excited.
And then they stop. Not because AI isn’t ready or because the ROI isn’t there. They stop because the path from “this makes sense” to “this is live in our environment” has real obstacles that no one in the board meeting is accounting for.
And so, AI in the revenue cycle has stopped short of a “tipping point.”
Why the AI Freeze Happens
The barriers show up consistently across health systems:
- IT misalignment. Almost always, there is a real gap between what the CEO is asking for and what IT is actually prioritizing. Revenue cycle leaders get caught in the middle. The mandate to implement AI in the revenue cycle exists, but the infrastructure pathway doesn’t.
- EHR dependency. A vendor that the health system already has tells them this is “on the roadmap.” It’s the path of least resistance, and in a resource-constrained environment, it’s easy to wait. The problem is that EHR-native AI tools are not designed to analyze payer patterns, dynamically score risk, or surface actionable, role-based intelligence across all payers and DRGs.
- Cost anxiety. Even when the math works, the upfront investment is a hard sell in an environment where margins have been shrinking. This is especially true when the person asking for ROI proof is also the person who controls the budget.
The Growing Need for AI in the Revenue Cycle
For many health systems, the barriers have stopped AI progress before it starts. However, the revenue cycle is genuinely the highest-ROI place to deploy AI in a health system today. Not because it’s flashy, but because the problem is measurable, there is plenty of data, and the financial impact is direct.
Administrative costs account for more than 40% of total hospital expenses. In 2025, hospitals spent nearly $18 billion just overturning claims denials, and the AHA estimates the total cost of chasing reimbursement from payers has reached $43 billion. Initial claim denials hit 11.8% in 2024, up from 10.2% a few years prior. Medicare Advantage denials jumped 55.7% between 2022 and 2023 alone.
Payers Have Been Using AI for Years
Payers have been using algorithmic review tools for years, including AI that flags, routes, and in some cases auto-denies claims at scale. There are documented instances of AI-driven denial systems rejecting hundreds of thousands of claims in weeks. Many revenue cycle teams are absorbing that volume manually.
That’s not a level playing field, and it’s not going to get more level by waiting.
What Revenue Cycle Leaders Can Actually Do Right Now
None of this requires an immediate enterprise-wide deployment. A few practical starting points:
- Establish a baseline before you need it. The hardest part of proving AI ROI is that most organizations can’t quantify what the problem costs them today. Before any vendor conversation, know your denial rate by payer, your overturn rate, your cost per rework, and your write-off volume. If you don’t have that data, that’s the first problem to solve.
- Separate the EHR conversation from the AI strategy conversation. Your EHR is not your AI strategy. It can be part of the picture, but waiting on an EHR roadmap item for revenue cycle AI is like waiting on your EMR to solve your staffing problem.
- Collaborate on a pilot. Most credible AI vendors will offer a pilot or beta program, or a sample on your own data. A pilot gives you real data to take to your CFO and your IT team. It also tells you whether the vendor’s claims hold up in your specific environment.
- Get IT involved early, not as a gatekeeper, but as a partner. The 51% of organizations that cite IT infrastructure limitations as their biggest obstacle are often dealing with a relationship problem as much as a technical one. Bringing IT into the evaluation phase, not just the implementation phase, helps change the dynamic.
The Gap Between AI Strategy and AI Execution
Revenue cycle leaders can absolutely go to a conference and talk about AI all day. The CFOs will too. But talk doesn’t prevent denials. And while health systems wait for the right moment to act, denial rates are climbing, payer AI is getting more sophisticated, and the gap between what’s owed and what’s collected keeps widening.
The health systems making progress aren’t the ones with the most polished AI strategy. They’re the ones that move past the sticking point. They started somewhere, measured, and made the next decision from there.
Interested in what your payer data is already telling you? Sift can help you establish the baseline.
