The Company She Wishes Didn’t Have to Exist
Anka Health’s Chief Revenue Officer came on the IACCGH Business Hour to explain how AI is fixing medical billing. She ended by making the case against her own industry’s necessity
By Somdatta Basu
HOUSTON – With about a minute left in the hour, Gayatri Garg said the thing that a Chief Revenue Officer is not supposed to say.
“I kind of wish that my company didn’t exist,” she told the hosts of Open Forum. “That this was not a space that had to exist.” Then she explained why it does. “The system feels so confusing. It’s the only place that I can think of, that you provide a service and you wait to get paid – if you get paid.”
It was an unusual close to an hour that had otherwise been a clear-eyed technical briefing, and it reframed everything that came before it. Garg, Chief Revenue Officer of Anka Health, had spent forty-five minutes on the IACCGH Business Hour on Open Forum FM 103.5 explaining how artificial intelligence is being applied to medical revenue cycle management. The argument underneath it was simpler and sharper: the billing apparatus American medicine has built is so baroque that an entire industry exists solely to navigate it – and the doctors in the studio knew exactly what she meant.
A number that stops the conversation
Host Dr. Subodh Bhuchar set the terms early, from the practitioner’s side. “We lost more money by not collecting money from the insurance companies that is legally due to us,” he said, “because of so many roadblocks and obstacles and authorizations – and they find every reason not to pay.” He described a friend who worked inside an insurance company and watched 10% of claims discarded automatically, on the assumption that most physicians would not appeal.
Garg had a figure of her own. For some of the practices Anka works with, collecting what they are owed costs one in every four dollars. A quarter of the revenue, spent on the act of retrieving the revenue.
That, she said, was what her father saw. He had built a business process outsourcing unit twenty-five years ago, backed by the Jindal Group, and over two decades of splitting time between India and the United States – and of listening to an uncle who practices gastroenterology in Chicago – he arrived at a conclusion. Physicians were staying up until midnight taking patient calls and were not being paid for the care they had already delivered. There ought to be a resolution to that.
The company’s name came, fittingly, from an AI query. Garg wanted something that would place first alphabetically in conference brochures and was not already taken. It also happens to mean prosperity in ancient Egyptian.
Three generations of a car
Asked what AI actually does inside a billing system, Garg reached for an analogy that landed better than any product description could.
When she arrived in the United States eighteen years ago, having learned to drive on a stick shift in India, she watched her cousin drive a Honda with one leg tucked beneath her, cruise control set, the car quietly slowing itself when traffic closed ahead. “I was mind blown,” she said. That was version one.
Version two came years later: lane assistance, automatic braking, a car that corrected small errors on its own. Version three is a Tesla – self-driving, and yet, she stressed, hands still very much on the wheel, the system alerting the driver the moment their eyes leave the road.
Revenue cycle AI, she argued, sits across all three. And her COO supplied the line that keeps expectations honest: what the technology delivers today is not artificial intelligence so much as “almost intelligent.” It cannot function if the underlying data and processes are wrong. “For me to come in and say AI has solved it – it is not. We still very much need hands on wheel.”
What it actually does
The mechanics, as she laid them out, are less glamorous than the marketing and considerably more useful.
Anka plugs into existing electronic health record systems – eCW, Epic, and others – without replacing them. It ingests a practice’s historical data, then looks for anomalies: repeated denials, repeated rejections, patterns that indicate a fixable process rather than a legitimate refusal. Where a pattern emerges, the fix is programmed into the practice’s system as an edit, not into the AI. “It’s teaching the entire process,” she said.
Once that groundwork is laid, the AI takes on the repetitive work: pulling medical records for documentation requests, routing them to the right payer, and – over time – autonomous resolution. If a payer suddenly demands prior authorization for a routine evaluation code, the system can check both that patient’s history and the practice’s full claims history and answer: you have never required this before.
The asymmetry she is trying to correct is stark. “On the one hand we have insurance companies who have a ton of technology and AI,” she said. “And then on the physician side you have a human being and a fax machine.”
The company’s data set runs back roughly ten years and holds about a billion records. Even so, every client takes three to six months to understand properly. “Every single organization we work with is a little bit of a snowflake.”
The typical result: a minimum 10% lift in collections.
The one question to ask yourself
When Dr. Bhuchar asked what she would tell a physician five years into practice who suspects something is wrong but cannot name it, Garg narrowed it to a single diagnostic.
Do you know your denial rate?
If you cannot answer, she said, that is the finding – it means your systems are not surfacing enough data for you to have visibility at all. And if you can answer and the number is above 5%, that is the place to start asking why. Some denials cannot be prevented. Many can.
She was equally direct about pricing, in a way that cut against her own industry. Most revenue cycle firms adopting AI charge heavy implementation costs while keeping their rates unchanged. Anka’s position is the opposite: if technology means fewer people doing the work, the client should pay less. “You should see the benefit of that.”
Rural Texas, and a closing hospital every few years
The conversation turned to rural health care, where Garg’s numbers were the afternoon’s most sobering.
In the 1960s, Texas had roughly 300 rural and community hospitals. Today there are 160. Twelve are at immediate risk of closure; another seventy face future risk.
Anka works with two rural Texas hospitals and physician practices scattered across the state. Garg pointed to genuine momentum behind the Rural Health Transformation program, noting that she had watched Governor Abbott sign related legislation at a conference of the Texas Organization of Rural and Community Hospitals – and she credited the effort as bipartisan, naming Representative Ann Johnson and Senator César Blanco alongside the Governor. “There’s a burning platform,” she said.
On the question everyone asks
Two callers – Daniel and Vikram – pressed the bigger question: is AI a threat to humanity?
Garg declined to overclaim. “I’m the Chief Revenue Officer. I’m certainly not the head of product – for a very clear reason.” But she offered a distinction worth keeping. Anka’s models are proprietary and live inside its own system; they are not open, not fed into a general-purpose engine, and the application is deliberately narrow and heavily monitored.
Her broader view was almost bureaucratic, and better for it. The models themselves are not the danger. The infrastructure built around them is the question – guardrails of the kind the FDA provides elsewhere. “It’s a productivity tool. It’s a way for us to do something faster.”
She had a personal example of that, too. Last year her nine-year-old daughter had ear surgery, and Garg – not a physician – fed background information into an AI tool. Not to diagnose anything, but for something more modest and more useful: which five questions to ask when she got into the room. “I know doctors also don’t have time. You’re running from one patient to the next. So how do I show up informed?”
Pajama notes
The doctors in the studio had one more request. They call it the pajama note – the charting that follows you home and eats the evening. Dr. Bhuchar was blunt about primary care documentation: patient attribution, labs, pharmacies, referrals and confirming they came back, imaging. “It’s bad news.”
Garg pointed toward ambient documentation tools rather than pitching her own product. It was characteristic of the hour. Asked repeatedly to claim more territory, she kept declining – on clinical AI (“not my forte, I’ll be very clear”), on functionality already built into a client’s EHR (“you don’t want to pay me for something that is already in your system”), and on the question of whether her technology eliminates billing staff.
On that last point she was thoughtful rather than reassuring. The goal is for AI to be a force multiplier in an industry where there is a perennial shortage of workforce. It is meant to absorb what is repetitive and manual, freeing staff for work that requires a person – talking to patients, working credentialing, negotiating with payers at the project level. She compared it to consultants who once drew charts by hand with a ruler and pencil, and who now analyze the data instead. The ruler and pencil did not put the consultants out of work. It changed what the work was. That, more or less, is the whole pitch – and it is a modest one by the standards of an industry that has spent two years promising to end medical billing as a problem.
Which is why the closing minute mattered. A Chief Revenue Officer who wishes her market did not exist is not being coy. She is describing the market accurately: a system in which the people who deliver care are the last to be paid and the least equipped to collect, and in which a company like hers is the workaround rather than the fix. Anka’s bet is that the workaround can be made cheap enough, and transparent enough, that practices stop treating a quarter of their revenue as the cost of doing business.
Until then, the advice stands. Find out your denial rate. If it is above 5%, ask why. And if nobody can tell you the number, you already have your answer.
The IACCGH Business Hour airs on Open Forum FM 103.5. Gayatri Garg’s appearance was recorded September 19, 2026.
About Anka Health
Anka Health is an AI-native revenue cycle management platform for mid-market healthcare providers, including physician practices and rural and community hospitals. Backed by the Jindal Group, Anka integrates with existing EHR systems to identify and correct the root causes of denials, automate repetitive billing work, and reduce the cost of collection. Anka is headquartered in Houston. More at ankahealth.ai.