Health

Prosper's $108 billion uncollected-care claim shrinks by half under its own no-show and denial rates

The a16z-backed voice-agent vendor says specialty practices leave $125,000 per physician uncollected each year. Run the report's own no-show, denial and collection rates back through its flagship example and the number shrinks by roughly half, and most of what remains is a transfer from payers and patients, not new money.

In this storya16z
Richard TangRichard Tang · 2 min read
Share
A receptionist with a headset at a clinic front desk, the calls Prosper AI's voice agents aim to answer
1 / 5Slide 1 of 5
Prosper AI's pitch rests on the front-desk call: the work its voice agents do for specialty practices.

Prosper AI, the a16z-backed voice-agent vendor for specialty practices, published research on October 13 claiming US dermatology, ophthalmology, GI, OB-GYN, primary care and orthopedic practices leave $108 billion a year uncollected, about $125,000 per physician across four levers 1. Its investor a16z amplified the number the same day 2. Run the report's own rates back through its own flagship example, and the headline does not survive.

The flagship example shrinks under the report's own rates

The flagship is Piedmont Plastic Surgery and Dermatology in North Carolina: 37 providers, 17% more appointments booked in August year over year, 3,315 extra bookings the report values at $1.3 million using MGMA's $392 dermatology visit 2. But the same report measures an 18.8% average no-show rate, a 10% first-submission denial rate, and practices collecting only 39% of what patients owe at the visit 2.

Apply those to the extra bookings and the revenue that survives to collection is roughly $740,000 to $950,000, not $1.3 million. The $108 billion is four best cases added up with no netting between them.

Calls are the biggest lever; the other three are money already owed

$0$20,000$40,000$60,000Calls and schedulingRe-engaged patientsPrevented denialsCollected balances$14,100$44,800
Data
Per physician per year
Calls and scheduling$44,800
Re-engaged patients$37,400
Prevented denials$28,500
Collected balances$14,100
Prosper AI's four revenue levers as stated per physician per year, in US dollars: answered calls and filled schedules, re-engaged patients, prevented denials, collected patient balances. Source: Prosper AI Growth Report, October 13, 2026 [1].1

Most of the four levers is money already owed

Three of the four levers, prevented denials, collected balances and re-engaged lapsed patients, are money already owed to someone 1. Practices control them, and that is the honest part of the pitch: answering every call and verifying benefits before a visit is real work, and Prosper's Piedmont result shows the capability is live 13.

But in a 2027 where Medicare Advantage moves 3.2 million enrollees into new plans, raises average drug-plan deductibles 23% and lifts out-of-pocket maximums 9% 5, while a 2.5% CMS wRVU haircut on roughly 7,700 codes lands on procedural specialties on January 1 4 and the temporary 2.5% Medicare boost dies at year end 1, most of that $125,000 is a transfer from payers and patients to practices, not growth anyone is creating.

The gap is real, and so far AI has not closed it

The profession's own new data says the gap is real but unclosed: MGMA's latest compensation report finds only 46% of practice leaders say AI made providers more productive 4. Several said AI cut documentation time without adding visits or revenue.

Prosper has raised a $30 million Series A led by a16z and sits in 100+ groups 6. The lesson for investors: discount the market-size slide to its own no-show math, and pay for the builder who closes the collection gap, which Prosper is showing it can do, one booked call at a time.

Get Health by email, free

AI in medicine, pharma and biotech. In our daily email, every weekday.

Reader comments

0 comments

    More in Health

    All in this topic →

    Health

    OpenEvidence's prior-auth letters no better than ChatGPT's in Mayo test; $15 billion buys its distribution

    A Mayo Clinic head-to-head and a Nature Medicine study put the quality of OpenEvidence's answers on the record in one week: good enough to use, no longer uniquely good. At $15 billion, investors are paying for distribution, and distribution now has company.

    Health

    Eli Lilly's 79% ovarian cancer response rate came from a dose its Phase 3 won't use

    Lilly's headline response rate at ESMO was measured at 4 mg/kg; the Phase 3 runs 3 mg/kg, and the combination result at that dose is unpublished. A day earlier, Genmab posted a positive Phase 3 on the same payload class.

    Health

    Molina's 10-Q moves its two biggest Medicaid wins past the 2027 recovery the Street is pricing

    The 10-Q moves the Georgia and Texas wins to late 2027 or 2028 while the Street raises targets on 2027 rates; the Florida children's contract is guided to a $1.50-per-share quarterly loss and about 700 autism providers say they are unpaid. The recovery-year revenue arrives after the recovery year.

    Health

    Intuitive Surgical to pay $450 million to settle hospitals' antitrust suit; revenue per procedure falls 3-5%

    Intuitive settled the hospitals' instrument-servicing antitrust case for $450 million, then disclosed its own 3-5% per-procedure price cut for 2027. The recurring revenue that carries 84% of sales now faces cheaper instruments on two fronts, just as US procedure growth slows to 10%.

    Health

    A compounded Zepbound recall hands Eli Lilly its exhibit as a fifth of US GLP-1 users take copies

    Apothecary Pharma could not guarantee the safety of 22 lots of compounded tirzepatide. With roughly a fifth of U.S. GLP-1 users on compounded copies, the October 29 and November 4 earnings prints will price what the copycat channel costs Lilly and Novo.

    Health

    Veeva sells AI flat, and IQVIA's seven-month clock starts the test

    Veeva priced Vault AI at 10 percent of a subscription with unlimited usage, betting model costs fall faster than adoption, while IQVIA's seven-month rollout puts a clock on the standardization it needs first.

    Get the newsletter

    The day's AI and market news, in your inbox every weekday.