top of page
Search

June 6, 2026: Healthcare AI Just Posted Real Numbers, and the Rest of the Organization Is Still Catching Up

  • Writer: James Sale
    James Sale
  • Jun 6
  • 5 min read

Three stories that landed this week come from completely different functions. Revenue cycle teams at hospitals are posting real numbers. WTW is handing C-suite leaders a diagnostic playbook for workforce redesign. Engineering teams are discovering that the code their AI tools write is often functionally correct and structurally dangerous. And a fresh CompTIA report shows that most organizations still haven't built the training infrastructure to support any of it.


The through-line is not that AI is advancing. It is that deploying AI and actually being ready for what it produces are two different problems, and most organizations are more invested in the first than the second.


Agentic AI Just Posted Its First Real Healthcare Report Card

At HIMSS 2026, the largest health IT conference of the year, the headline was not a demo. It was a number. Waystar reports $15 billion in prevented denials and 90% reductions in appeal workflow time since deploying agentic AI across revenue cycle management. FinThrive reports 1.1% underpayment recovery worth nearly $1 million in three months through agent-driven analysis.


Those are vendor-reported figures, and they should be read with appropriate skepticism about methodology and selection bias. But their scale and specificity are worth paying attention to. The Revele MD recap describes a conference where agentic AI, meaning AI that takes autonomous action across multi-step workflows rather than surfacing recommendations for humans to act on, moved from demo stage into something you could point to on a P&L.


Vendors including Waystar, FinThrive, XiFin, Solventum, Inovalon, and Innovaccer all announced agentic capabilities focused on prior authorizations, denials, appeals, and coding. Results cited include 42% reductions in prior authorization turnaround time alongside the appeal workflow gains.


For practice managers and revenue cycle leaders, the question is no longer whether this technology works in principle. It is what implementation actually requires. Clean data going in, change management for the billing team, and a clear owner for the redesigned workflow. Organizations with fragmented payer data and legacy system constraints will see much longer timelines before those numbers translate.


WTW Gives the C-Suite a Framework Instead of a Guess

Knowing that AI could change your workforce and knowing where to start are two entirely different problems. Most organizations are still operating somewhere between the two.


WTW's new AI Workforce Transformation offering addresses that gap directly. The proposition is built on WTW's proprietary data on jobs, skills, and work processes, and it includes two diagnostic tools: WorkVue Agent, which maps automation potential by job across an organization, and ChangeVue, which identifies where adoption is most feasible given current readiness.


In the launch announcement, Julie Gebauer described the offering as giving "C-suite leaders the evidence they need to add AI where it drives the most productivity and growth, and to move faster than competitors who are still guessing."


That framing is pointed for a reason. A lot of AI workforce strategy right now is still guessing. Organizations run pilots based on vendor enthusiasm rather than a structured view of which roles have the highest automation potential and which teams can actually absorb the change. The diagnostic layer WTW is offering should, in principle, come before deployment decisions, not after them.


If your current AI strategy skips straight to tooling without an explicit readiness assessment, that gap tends to surface six months into a rollout when adoption stalls.


The Code Is Working. The Security Isn't.

AI coding agents like Cursor and Claude Code are now a routine part of many engineering workflows. The problem, which Endor Labs' AURI platform is designed to address, is that AI-generated code is often functionally correct and structurally insecure. Real-world incidents cited include dropped databases and production wipeouts.


A parallel piece from Snyk, published the same week, makes a complementary point: simply instructing large language models not to include vulnerabilities does not reliably work, and traditional code review approaches were not designed for the volume and patterns that AI-generated code introduces.


AURI functions as a security intelligence layer sitting between the AI coding agent and production, addressing vulnerabilities at generation time rather than catching them during post-commit review. The concept of shifting security left, meaning addressing it earlier in the development cycle, is not new. Applying it specifically to the AI code generation point is, and the urgency is growing as AI-assisted code volume scales.


For engineering leaders, the honest question is whether your security review process was built for teams writing code by hand or for the pace and volume of what AI agents now produce. Those require materially different approaches.


The Budget Is There. The Training Infrastructure Isn't.

Connecting all of this is a gap that CompTIA's 7th annual Workforce and Learning Trends 2026 report makes concrete. Per the report, 62% of HR professionals and IT leaders expect AI training budgets to increase in the next year. 83% expect skill development to have high or moderate impact on employee morale and engagement. Job role-based training ranks as the top preferred format.


That sounds like momentum. But read alongside a separate 2026 Economist Impact study of 639 decision-makers, which found that only 16% of organizations offer structured internal AI training despite nearly all claiming to take some action on AI, and the picture gets more complicated. Intention and infrastructure are not the same thing.


The gap between organizations with committed AI training budgets and those with actual structured programs, dedicated curriculum, and measurable proficiency benchmarks, is where most organizations quietly sit. The CompTIA budget signal is a window. Organizations that convert it into role-specific, structured programs in the next 12 months will be in a materially stronger position than those treating "AI training" as a vendor demo day.


The human reality underneath all three of today's stories is the same. People across healthcare administration, HR, software engineering, and workforce development are being asked to absorb significant changes to how their work gets done. The ones with clear frameworks, honest readiness assessments, and structured support are navigating it. The ones running on enthusiasm alone tend to find out the hard way when the first rollout hits friction.


Worth Acting On

Map your revenue cycle AI readiness against your own data quality before vendor benchmarks. Before committing to agentic AI for prior auth or denials, assess payer data cleanliness and EHR integration first. The headline numbers assume clean inputs that many organizations do not yet have.


Run a workforce readiness assessment before selecting AI platforms. WTW's launch underscores what organizations routinely skip: knowing which roles have genuine automation potential and which teams can absorb the change. That diagnostic should come before platform decisions, not after.


Audit whether your code security process was designed for AI-generated volume. If your review cadence was built for human-written code, it is likely underprovisioned for what AI coding agents now produce. The exposure is not always visible until something breaks in production.


Convert rising AI training budgets into structured, role-specific curriculum. CompTIA shows budget intent is growing. Most organizations still lack structured programs to match. The difference between those two things is accountability: who owns the curriculum, who measures proficiency, and what role-specific outcomes are expected.


The harder question: If you removed vendor-reported outcomes from your organization's AI business case, what measurable evidence from your own operations would remain to justify the next phase of investment?


If you want to stay current on how AI is reshaping workforce strategy, healthcare operations, and engineering workflows, and what it means for the people navigating those changes, Agenticism is where those stories live every day. For the curated weekly, monthly, and quarterly digest delivered to your inbox, subscribe at Agenticism on Substack.


Sources

Recent Posts

See All
bottom of page