In fifth grade, Bakhtawer Baloch spent 100 Pakistani rupees, less than a dollar, on stickers. She sold them and walked home with a thousand rupees, ten times her money. "I came back home, and
In fifth grade, Bakhtawer Baloch spent 100 Pakistani rupees, less than a dollar, on stickers.
She sold them and walked home with a thousand rupees, ten times her money.
"I came back home, and I think I realised I am going to be doing this all my life," she said in the latest episode of Against The Odds, TheStreet Roundtable's interview show hosted by Tina Das. Each episode features a founder who built something real from an unlikely starting point.
Two decades later, the 31-year-old is in New York, pursuing her second master's degree at NYU and running Clairdoc, a nine-month-old startup that uses AI to predict whether an insurer will pay a medical claim and, when it will not, exactly why.
Bakhtawer Baloch is the founder of Caliredoc, an AI startup that helps doctors reduce their billing denial rate
Through special arrangement/Bakhtawer Baloch
Baloch grew up in Pakistan, the daughter of two entrepreneurs. Her father, an engineer, built a real estate business from scratch. Her mother ran a successful salon he had helped her set up.
He had married outside his community, and he insisted all three of his children get the best education available, regardless of gender.
In Baloch families, that was unusual.
For context, Baloch are one of Pakistan's largest ethnic groups, a tribal people whose homeland spans Pakistan, Iran and Afghanistan, and their society is among the country's most conservative. In Balochistan, the province named after them, fewer than a third of women can read, the lowest female literacy rate in Pakistan, and nearly half of women marry before 18. Even within her extended family, she watched women steered toward early marriage and away from school.
She went the other way.
Baloch topped Pakistan's chartered accountancy exams, finishing in three years, and joined Deloitte.
Related: Mace Ai founders on what OpenAI's Hugging Face breach means for businesses
Baloch earned a master's from the Isenberg School of Management at UMass Amherst with a 3.9 GPA, then spent a year at Deloitte in the U.S.
In 2020, she started freelancing and built a profitable U.S. medical billing and consulting business, growing it to a team of 15 to 20 people over four years.
She moved to New York in 2024 for a second master's at NYU. Along the way, she worked on lung cancer research at NYU Langone Health, helping the team run its pipelines on high-performance computers, and later worked in product management at TikTok.
Claire Doc combines machine learning paired with the rules that govern U.S. billing. Representational image. Source:Getty Images
The idea for Clairdoc came from a hospital room. While Baloch was completing her first master's in data science and AI, her father, a kidney transplant patient, visited the U.S. and his blood pressure spiked.
At one of the city's best hospitals, she watched staff take notes by hand and new nurses arrive with no access to his records.
"Even though America is so far ahead in terms of technology... you walk in, they're writing things down with their pen," she said.
Having worked the billing side, she knew what came next. Paperwork errors turn into denied claims, and denied claims turn into bills for patients.
Every visit to an American doctor generates a claim to the patient's insurer. If the claim is incomplete or coded wrong, the insurer can reject it, and rejections are common.
Insurers on HealthCare.gov denied 19% of in-network claims in 2024, according to a KFF analysis of federal data, with rates ranging from 3% to 36% depending on the insurer. Of roughly 451 million in-network claims filed, about 85 million were denied.
Only 5% of denials were about medical necessity, KFF found, while a quarter came down to administrative reasons. Consumers appealed fewer than 1% of denied claims, and insurers upheld 66% of those appeals.
Clinics often do not fight back either. Industry research cited by the Medical Group Management Association puts the average cost of reworking a denied claim at $25.20, and most denied claims are never reworked.
Related: GTA 6 may be the last blockbuster built without AI
Researchers writing in JAMA estimated billing and coding waste alone costs the U.S. health system about $248 billion a year. Health Affairs puts the broader administrative waste problem at up to $570 billion.
The pain is sharpest in the clinics Baloch targets. Outpatient specialists such as gastroenterologists and dermatologists bring in roughly $1 million to $5 million a year but run on one or two billers.
"They have the same billing complexity of a hospital with a small team," she said.
Then there is underpayment. If a clinic files a claim for $1,000 and receives $800, most billers simply move on, she said. They have hundreds of claims waiting and 10 to 15 minutes for each one.
When a denial sticks, the patient often ends up paying.
"It's a lose-lose situation for both the patient and the doctor," Baloch said, "but it is a win-win situation for the insurance company at all times."
Those hit hardest, she added, are lower-income patients, who are more likely to visit clinics with stretched billing staff.
The numbers behind U.S. claim denials, and where Clairdoc wants to take them. Source: TheStreet RoundtableClairdoc reviews claims before they go out and flags the ones likely to be denied, along with the reason. It also audits past claims to find money insurers still owe.
The technology is called neurosymbolic AI. In plain terms, it pairs machine learning with the written rules that govern U.S. medical billing. A purely probabilistic black-box model, she said, cannot explain why a claim fails.
"And if you don't know how to fix a problem, how are you going to fix it?"
Many denials hinge on an insurer deciding a patient was less complex than the doctor claimed. Baloch said her team published research at ICML, a major machine learning conference, on patient complexity, and built from it a comorbidity index that predicts how insurers will read a case and tells physicians what their notes must document to get paid.
The target is to bring clinics' denial rates from around 15% to under 8%.
"Our goal is to make sure that the doctor gets paid the first time and correctly and the patient can get the best of care," she said.
The company is part of Google for Startups and NVIDIA's Inception program, and it came through NYU's startup incubator, where Baloch was named a Female Founders Fellow this fall. She has also pitched Clairdoc at NYU's Demo Day.
Venture funding for AI billing startups is surging, Fierce Healthcare reported in July, as Candid Health raised a $120 million Series D led by Sixth Street Growth. That brings Candid's total to more than $219 million. The company says more than 200 healthcare organizations use its software and that it processes about $7 billion in claims a year.
Adonis, a New York company, raised a $40 million Series C in March and counts Mount Sinai among its users. AKASA has raised more than $200 million. The incumbents are moving too. Waystar, which serves about 30,000 clients, closed its acquisition of Iodine Software on Oct. 1, 2025, adding AI to a platform that handles more than $1.8 trillion in claims a year.
Most of these companies sell to hospitals and large provider groups. Clairdoc is aiming at small specialty clinics instead. Whether that is a gap the big players have left open or a market too small to build a business on is the open question.
CB Insights studied 431 venture-backed startups across all sectors that shut down since 2023. It found 43% cited poor product-market fit, meaning they never found enough customers who wanted what they built. The median company died 22 months after its last funding round.
Large companies struggle with AI too. An MIT study of generative AI projects, reported by Fortune, found only about 5% of pilot programs produced rapid revenue growth. The same research found that buying AI tools from specialist vendors worked about 67% of the time, a point in favor of vendors like Clairdoc.
Before building anything, Baloch went through NYU's startup incubator boot camp, then spent months on customer discovery, interviewing more than 100 doctors and 50 billers.
"Building now has become much easier," she said. "The most important thing for founders now is how to sell and what to sell."
Selling was the hard part. Healthcare is closed to newcomers, Baloch says, and integrating with electronic health record systems is expensive. So she started with her own network.
Her first customers were clinics already losing money to denials, she said, and one doctor let her test the tool on his past claims.
Today, Baloch says, the company works with more than 12 physicians across more than eight clinics, all in dermatology and gastroenterology.
Related: How blockchain's are turning AI's quiet expenses into businesses
