A quick Google search (or generative AI) will provide hundreds of statistics on the dire state of healthcare costs plaguing our country. Whether it is the fact that we spend nearly 5 times higher than neighboring Canada per capita in health care administrative costs or whether it means every quarter of a dollar spent is wasted, there are many scary statistics that we have become accustomed to reading.
This definitely wasn’t supposed to happen.
How did a system designed to heal become so broken and complex that it haunts us financially and clinically?
The untamed beast that keeps growing
Two centuries ago, Thomas Jefferson said that “the strength of a country lies in the health of its people.” He couldn’t have imagined a world in which seeking care requires deciphering 800-page insurance decisions before seeing his provider. Until a few decades ago, a physician using pen and paper could treat patients without needing to log into (or complete) multiple payer portals (or prior authorization forms) or justify each procedure in triplicate. This simplicity now seems terribly outdated. The biggest question is undoubtedly: are we unable to keep “care” at the center of the “continuum of care”?
Today, doctors spend 17% of their working week (can be up to 10 hours depending on their healthcare setting) solely on administrative tasks. Prior authorization, billing, mapping, coding, data entry into EHRs, RCM audits, quality reporting and anything else that hasn’t metastasized. HITECH’s electronic record-keeping campaign, while well-intentioned, has resulted in siled systems and mandatory documentation that demand more time at home, eroding both clinical presence and work-life balance.
But is technology the real problem? Or processes?
Maybe none of that. Each part of this system works perfectly independently, it is when they come to fruition that the real disaster occurs.
From hospital leaders to frontline providers to policymakers and payers, everyone wants to solve this problem. No one set out to build a monster. Every new rule, digital upgrade, coding standard, or quality measure was created with the best intentions. Yet over the decades we have unwittingly built up a bureaucracy too complex to reform. Providers entered medicine to heal, not to navigate mazes. We’re all stuck, not because of malice, but because the system grew in pieces that no one knew how to stitch back together.
The new entrant: AI
There has been a lot of talk about AI in healthcare, as in any other sector. The hype is bigger than anything seen since the dot com boom. And it’s completely real. The enormous US administration bill of $353 billion is already being reduced to nothing, with experts forecasting a $168 billion reduction in annual administrative costs through the adoption of AI tools in the near future. Another analysis by CAQH discovered approximately $20 billion in near-term savings by moving from manual to electronic workflows around eligibility, pre-authorization and payment reconciliation, among others. The rulebook was expanded from around 14,000 ICD-9 codes to almost 70,000 in ICD-10, so an increase in edge cases was created and refusals as well as reworks were multiplied. Of more than 1,000 healthcare finance, technology and RCM executives surveyed by Black Book Market Research, 83% reported that AI-powered automation reduced claim denials by at least 10% in the first 6 months of implementation.
Real-world excellence and anticipated possibilities
With open LLMs running on modern GPU stacks, we are now able to reason about complex claims in real time. By bootstrapping high-quality examples and using reinforcement learning to tune these models with optimal reward strategies, then connecting them into pipelines that surface payer policy changes as they happen, we finally see true adaptability emerge. These developments distinguish this wave from the big data frenzy of the 2010s. Although data fragmentation, strict privacy regulations, and aging hosts continue to complicate the situation, I sense this momentum growing every day. RCM is truly the first administrative area where AI actually reduces workload rather than adding more clicks. This phenomenon will continue to grow if incentives remain consistent and algorithms are paired with human listeners.
Denial prevention is now proactive; according to a study out of 102 million hospitalizations, 86% of denials could have been avoided, a use case now being solved at scale by health tech evangelists. Many of these are flagged by contemporary classifiers at the time of claim creation, allowing staff to correct eligibility or modifiers before money is at risk. Regarding documentation, AHIMA has do Real-time question responses and agreement evaluate a key CDI KPI, highlighting how point-of-care nudges reduce DNFB delays and improve the accuracy of safety indicators.
I see AI as the linchpin of innovation; we are already testing CDI engines that map V28 risk in real time, eligibility checks that complete while the patient is still present, and LLM agents that compose appeal records. Contactless RCM technology is the logical conclusion; it reconciles charge entry, documentation, and payment integrity before a claim leaves the EHR, giving doctors time back and providing executives with a revenue stream that finally works the way the spreadsheets dictate.
Disclaimer and way forward
The only way for our health system to emerge from the administrative chaos in which it currently finds itself is through the grassroots, through a coalition at all levels. This is why people like me are optimistic about AI. Rethinking workflows, simplifying RCM processes, removing code, realigning incentives: all of this can be facilitated by targeted AI.
Seeing good, hard-working healthcare professionals in a helpless situation is always upsetting, but I am confident that newer technologies and more sophisticated efforts will allow us to slowly return to a patient-centered, physician-responsive system. My only warning is that one-off solutions are ineffective. A comprehensive platform strategy is the best course of action if AI wants to make a dent. Otherwise, every touchpoint will continue to be disconnected, as before.
I’m sure the bump is around the corner.
About Dushyant Mishra
Dushyant Mishra is the co-founder and CEO of Quick complaintswhere he combines deep health technology expertise with a pragmatic approach to AI through his decade-long experience solving financial flaws in the US healthcare system. RapidClaims partners with health systems, hospitals, IDNs and physician groups to shift revenue cycle work from reactive remediation to proactive prevention. Through seamless, human-centered automation, the New York-based company eliminates denials, ensures legitimate payment, and gives healthcare teams their time and peace of mind back. As CEO, his leadership combines product rigor with an operator’s margin focus, earning the trust of healthcare providers who now rely on RapidClaims to recover millions in lost revenue each year.
