Frontera Chief Executive Amol Deshpande says consumption-based pricing and AI-driven efficiency, not billable hours, will define ABA’s next chapter.
Key Takeaways
- At Frontera, the access-to-care mission comes first, and the company puts its money where its mouth is. It was the first to offer free credits to any clinic taking on homeless or foster children, and it has stuck by that stance.
- ABA’s reliance on high-volume billable hours has become its central liability, with federal auditors flagging hundreds of millions of dollars in improper Medicaid payments and states from Nebraska to Indiana cutting rates and capping weekly service hours. Providers built to maximize authorized hours now face refunds, denials, and forced exits from entire markets.
- Frontera Health, the Denver AI startup led by Farmers Business Network co-founder Amol Deshpande, argues the billable-hour model doesn’t incentivize access and quality. Its assessment and diagnosis software is designed to cut clinician hours, a feature Deshpande says clinics first resisted but now increasingly treat as a competitive necessity.
- Rather than per-seat software licenses, Frontera sells “tokens” that clinics pay for only when they use them. Deshpande says the structure protects providers that shrink or leave markets and ties his company’s revenue to delivered value, unlike multiyear contracts that bill regardless of use. Further, Frontera must continue to innovate to satisfy customer needs, a fair tradeoff.
- Through a public partnership with the nonprofit Catalight, Frontera is also backing parent-mediated care for milder cases, a lower-hour model with a growing research base. Whether payers and providers come to reward efficiency over volume remains the open question that will decide whether the approach scales.
For most of the past decade, the business logic of applied behavior analysis was simple: more authorized hours meant more revenue. Medicaid expansion opened a fast-growing funding stream, private equity poured in, and a generation of providers scaled by filling schedules. That logic is now colliding with a wave of audits, rate cuts, and fraud allegations that have turned high-volume billing into the sector’s defining risk rather than its growth engine.
Amol Deshpande wants to sell clinics the opposite bet. The Chief Executive and co-founder of Frontera Health, a Denver startup that builds artificial intelligence tools for autism assessment, argues that the providers most likely to survive the current scrutiny are the ones billing fewer hours, not more. “You don’t just give 40 hours for every kid,” he said in an interview with Acuity Media Network. His company’s pitch, and its pricing, are built around that conviction.
Taking on the hardest cases is also part of the pitch. Access to care, Deshpande argues, is not just a statement or buzzword at Frontera; it is everything. The company announced months ago a program in which any clinic serving foster or homeless children receives free credits for those cases, meaning the tokens used on those assessments are free. A clinic that handles only foster or homeless cases can use Frontera entirely for free. The thinking is simple: most clinics chase billable hours, high rates, and the easiest cases, while Frontera is making a statement meant to encourage clinics to take on hard cases and do so profitably with new AI technology.
A Sector Built on Hours, Now Under the Microscope
Start with the numbers. Indiana’s Medicaid ABA spending climbed from $21 million in 2016 to $611 million in 2023, the steepest growth of any state program in the country. A federal Office of Inspector General audit of Indiana’s 2019 and 2020 payments identified at least $56 million in improper claims, with every sampled case containing at least one deficient line. A separate OIG review released in 2026 found $285.2 million in improper and potentially improper ABA payments in Colorado.
The states’ response has been blunt. Nebraska cut reimbursement rates by between 28% and 79%, depending on the service, and imposed a 30-hour weekly cap; within months, providers reported closing offices and dropping families. Indiana began a phased 10% rate reduction in April 2026 and restricted coverage to members under 21. North Carolina, where spending rose 423% in four years, attempted a 10% cut that was challenged in court, after which the governor ordered rates restored.
The pressure has landed hardest on the operators that grew fastest. Private equity firms acquired more than 500 therapy centers over the past decade, most between 2018 and 2022, and several of the largest chains are now navigating audits and payer disputes. ABA Centers of America, one of the sector’s fastest-growing companies, is facing a $19 million fraud counterclaim from a Massachusetts insurer that alleges aggressive out-of-network billing and pressure to maximize billable hours, claims the company disputes.
Frontera’s Argument: Efficiency Is the Product
Deshpande, who co-founded the agriculture technology company Farmers Business Network before turning to autism care after his son’s diagnosis, frames the reckoning as a market correction that favors efficiency. Frontera’s software uses AI to help clinicians produce assessment and diagnostic reports, work that he says can compress an eight-to-ten-hour task into roughly two to three hours with increased quality.
Founded in 2023 and backed by a $32 million seed round led by Lux Capital and Lightspeed Venture Partners, Frontera remains a small player relative to the chains it sells to, and its broadest claims about reshaping the economics of autism care are unproven at scale. What it has is a thesis aligned with where regulators are pushing: that the sector’s growth rested on a payment structure no longer politically or fiscally sustainable. The company also operates its own clinics, which Deshpande cites as proof that he is not asking customers to do anything he would not do.
That is a feature, in his telling, even though it lowers what a clinic can bill. “The path of AI isn’t displacing jobs or replacing clinicians,” he said. “It’s taking a clinician and leveraging their time.” The framing echoes a recurring argument among behavior analysts that AI will augment rather than replace the BCBA, though Frontera’s version carries a commercial edge: clinics that finish assessments faster, he argues, can clear waitlists and serve more children, rather than stretching billable hours per case.
When Frontera began selling its technology in 2025, Deshpande said, it wasn’t without resistance from clinic executives, who often objected that the tools would reduce their billable hours. His answer was to reframe the question: “Why aren’t you focused on getting more kids into care, not maximizing the billable allocation of assessment hours?” As audits and rate cuts spread through 2025 and into 2026, he says that pitch grew easier to make, with clinics increasingly treating efficiency as a condition of staying competitive.
Tokens Versus Seats
Frontera’s philosophy shows up most plainly in how it charges. Most practice management software in the sector is sold on a per-seat basis, with clinics paying an annual license fee for each user regardless of how heavily the tool is used. Deshpande describes the ERP players as data monopolies that prevent innovation and build moats around clinics’ data, locking in seats for years and then refusing to truly innovate, citing that very same “lock-in.” Frontera sells what he calls tokens instead: a clinic pays a set amount, roughly $50, each time it runs an assessment or reassessment, $100 for a diagnosis, and nothing when it does not. Clinics can stop anytime.
The distinction matters most when a clinic contracts or fails, he argues. A provider that bought hundreds of seats on a multiyear contract and then had to exit a state (as several have amid the rate cuts) would still owe for licenses it no longer uses. “The whole SaaS software game was to lock people into long-term contracts,” Deshpande said, contrasting it with consumption-based pricing that he says shifts risk back onto the vendor to keep delivering value. In a slow month with few assessments, a clinic pays little.
The model also functions as a hedge against the volatility now defining the sector. With Nebraska’s largest ABA provider having briefly threatened to leave Medicaid over rates and smaller clinics closing locations, Deshpande argues that providers should avoid fixed costs that assume perpetual growth. It is also, plainly, a sales position: Frontera’s revenue rises with usage, and “outcomes-based” is as much a market differentiator as a clinical philosophy.
The Parent-Mediated Bet
Frontera’s argument is not only about pricing. It extends to how therapy itself is delivered. In January 2026, the company announced a partnership with Catalight, a large nonprofit behavioral health network, to support what is known as parent-mediated care, in which trained caregivers deliver much of the intervention under clinical supervision. Catalight has championed the approach for mild-to-moderate cases, citing research suggesting it can be as effective as practitioner-led ABA for many children while requiring far fewer billed hours.
That position places Frontera on one side of a real clinical debate. Critics of the high-hour standard, including some providers, have likened the routine prescription of 30 to 40 hours a week to giving every patient the most intensive treatment regardless of need. Some of the country’s largest payers have begun making a version of that argument in public, questioning on earnings calls whether every child needs forty hours a week, and whether that intensity should hold for years. But families and advocates have also warned that rate cuts and hour caps, the same forces Deshpande sees as validating his model, can sever access for children who genuinely need intensive care, and several states have drawn backlash for moving too quickly.
Deshpande is careful to say parent-mediated care is not a universal answer. “There are a lot of cases for which that won’t work. Everyone knows that,” he said. The goal, he argues, is to match intensity to need instead of defaulting to the maximum. That depends on two things he does not control: whether payers build reimbursement that rewards calibration over volume, and whether providers can still make the math work as rates fall. Until that happens, the model Frontera is betting on remains exactly that, a bet.
