# LunaBill

Aiolum story: https://aiolum.com/story/lunabill/ · added 2026-11-05 · updated 2026-11-06

**In short.** LunaBill (AI voice agents for insurance claim follow-up calls, USA): US$428K live annual recurring revenue, November 2025. Evidence grade D: reported, not verified. Team: 3 founders in October 2025; 4 people on the Y Combinator page from November 2025 to November 2026. Launched July 2025; still running.

AI voice agents for insurance claim follow-up calls · USA · AI-native · team of 2 to 10 · still running · group: Clinic and practice admin

- Outcome: US$428K live annual recurring revenue, November 2025 (evidence grade of the headline number: D; headline in US dollars: 428000)
- Launch: July 2025; months from launch to the outcome: 4
- Team: 3 founders in October 2025; 4 people on the Y Combinator page from November 2025 to November 2026
- Outside funding: Y Combinator, Fall 2025 batch; amount not published in the sources read
- Main factor: Domain expertise
- Meters, 1 to 4: evidence strength 1, copyable pattern 1, capital needed 2
- Tags: AI-native, Small team, Funded, Running, B2B, Healthcare billing

Links:

- Visit lunabill.com: https://lunabill.com/

**The product.** Hospital and clinic billing teams spend much of their day on the phone with insurance companies to ask why a claim is not paid. LunaBill's AI voice agents make these calls: they handle claim-status checks, appeals follow-ups and reprocessing requests, connect to the customer's billing software to pick the claims that need work, and write a post-call summary with the next steps into the team's workflow.

**Why it worked.** The founders had worked inside healthcare billing before, so they knew that insurance follow-up calls are a large, repetitive and expensive part of the work. A voice model can now hold this kind of short, scripted business conversation. The company sells to health systems and billing firms through pilots, and the founders said in November 2025 that every pilot had become a paying customer.

Evidence grades: A = disclosed by a public company, a filing or a payment processor, or verified by one; B = two or more independent reports agree, or a journalist confirmed it; C = a founder claim that no one else verified; D = reported, not verified: one founder claim or a third-party estimate with no second source; E = no revenue figure found, so the headline is a non-money fact and the other facts carry their own grades; calc = computed on this page from graded facts.

## Summary

LunaBill sells AI voice agents to healthcare billing teams in the United States. An AI voice agent is a program that speaks and listens on a phone line: it turns the other person's speech into text, decides what to say with a language model, and answers with a synthetic voice. LunaBill's agents call insurance companies (payers, the companies that pay medical claims) to ask about the status of a claim, to get a stalled claim reprocessed, and to follow up on appeals. This work is part of revenue cycle management (RCM), the administrative process by which a hospital or clinic gets paid for care it has already given.

The founders are Suhail Parry (CEO), who ran billing operations at a billing company with more than 200 healthcare clients, David Day (CTO), who founded an AI medical scribe company in Taiwan, and Yash Raj (COO), who worked on clinical AI research at Mayo Clinic. They met at the Berkeley AI Hackathon. The product launched in July 2025, and the company joined Y Combinator's Fall 2025 batch. On the company's Y Combinator page, first archived with this text on 23 November 2025, the founders wrote that they had US$764,000 in contracted annual recurring revenue (ARR, yearly subscription revenue under signed contracts), of which US$428,000 was live revenue already being billed. That founder statement is the headline of this page.

The headline rests on one founder statement on one date, so it is grade D: reported, not verified. No payment processor, filing or journalist confirms it. The company's own website could not be read for this page (it answers automated readers with a bot check), so its current product details and prices are not published on a page we could read. The page also shows three failed companies that tried to automate hospital administration before voice models existed.

## Pricing

LunaBill prices: not published on a page we could read (the company site answers automated readers with a bot check; checked 5 November 2026)

| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Pilot, then contract | not published on a page we could read | not published on a page we could read | Pilot first, per the Y Combinator page |

The Y Combinator company page and the launch post of November 2025 give no prices. The founders wrote that customers start with pilots and that every pilot became a paying customer.

## How much revenue does LunaBill make? The numbers, graded

A four-person Y Combinator company that reported US$428,000 of live annual recurring revenue four months after launch, from AI calls to insurance companies

| Fact | Value | Grade | Note | Sources |
|---|---|---|---|---|
| Live annual recurring revenue, November 2025 | US$428,000 | D | Founder statement on the Y Combinator company page: 'we are now at $764K in contracted ARR, of which $428K is live revenue'. The text has no date of its own. It is missing from the Internet Archive copy of 19 October 2025 and present in the copy of 23 November 2025, so this page dates it November 2025. The page still shows the same figure on 5 November 2026. | [1, 2, 3] |
| Contracted annual recurring revenue, November 2025 | US$764,000 | D | Same founder statement. The founders wrote that the rest would go live by January 2026. No later revenue figure was found. | [1, 2] |
| Contracted revenue not yet live | US$336,000 a year | calc | Calculated: 764,000 minus 428,000. Live revenue was 56% of the contracted total. | [1] |
| Live revenue per month | about US$35,667 | calc | Calculated: 428,000 divided by 12. | [1] |
| Launch | July 2025 | C | Founders: 'Since our launch in July'. Y Combinator lists the company as founded in 2025. | [1] |
| Time from launch to the headline figure | 4 months | calc | Calculated: July 2025 to November 2025. | [1, 2] |
| Team size | 3 in October 2025, 4 from November 2025 | B | Y Combinator directory field 'team size' in the archived copies of 7 and 19 October 2025 (3) and 23 November 2025 (4), and on the live page of 5 November 2026 (4). | [1, 2, 3] |
| Founders | Suhail Parry (CEO), David Day (CTO), Yash Raj (COO) | B | Y Combinator lists the three as founders. The launch post gives the titles. | [1, 4] |
| Outside funding | Y Combinator, Fall 2025 batch | B | The amount is not published in the sources read. | [1] |
| Calls automated | 50,000+ (company page), 60,000+ (launch post of 11 November 2025) | C | Company statements. The lower count on the company page is one reason to date that text before or near the launch post. | [1, 4] |
| Training data | more than 1.2 million call transcripts | C | Launch post: the agents are 'fine-tuned on 1.2M+ real call transcripts'. The base model and the voice provider are not named. | [4] |
| Pilots converted to paying customers | 100% | C | Founder statement on the Y Combinator page. The number of pilots is not given. | [1] |
| Concurrent agents per client | up to 100 | C | Launch post. | [4] |

How to read this table: a grade D value is a single founder statement with one date that no other source confirms. A grade C value is a statement by the company on its own pages. A grade B value is a fact that a third party records, such as the Y Combinator directory fields or an Internet Archive capture date. A "calc" value was computed on this page.

## When did LunaBill launch, and how did it grow? Timeline

From a hackathon team to an insurance-call product with a reported US$428,000 of live yearly revenue in four months. Each entry carries its grade.

- **Before 2025** (grade C [4]): The three founders meet at the Berkeley AI Hackathon, where they build an AI nurse. A few months later they team up to automate insurance phone calls.
- **July 2025** (grade C [1]): LunaBill launches its AI callers for billing teams.
- **7 and 19 October 2025** (grade B [3]): The Y Combinator directory lists LunaBill with 3 people and the line 'Automating Calls for Healthcare Billing Teams'. No revenue text yet.
- **11 November 2025** (grade C [4]): Launch post on Y Combinator: 60,000+ insurance calls automated, partnerships with health systems and three large billing companies, and a request for introductions to hospital finance leaders.
- **November 2025** (grade D [2]): The company page, archived on 23 November 2025, says the founders reported US$764,000 of contracted ARR, US$428,000 of it live, with 4 people.
- **5 November 2026** (grade B [1]): Y Combinator still lists the company as active with 4 people and shows the same revenue text. The company site answers automated readers with a bot check.

## Why did LunaBill work?

### A narrow, repetitive phone conversation

A claim-status call follows a script: name the provider, give the claim number, ask for the status, ask what is missing, get a reference number. That narrow scope is what a voice model handles well. The launch post says the agents were fine-tuned on more than 1.2 million real call transcripts, so they learn the phrases and the menu logic of each insurer.

### Founders from inside the industry

Suhail Parry ran billing operations for a billing company with more than 200 healthcare clients before LunaBill. That gave the team the problem, the vocabulary of claim denials and, according to the launch post, the first contacts with health systems and billing firms.

### Pilot, then contract

The founders wrote on the Y Combinator page that they had converted 100% of pilots into paying customers, and that customers saw a tenfold increase in claims followed up per biller in the first week. A pilot on real claims shows the result in money the customer can count. The number of pilots is not given.

## How it was built

The sources say little about the technology. These details come from the Y Combinator launch post and company page, so they are grade C.

- Voice agents fine-tuned on more than 1.2 million call transcripts of insurance calls. The sources do not name the base language model, the speech-to-text service, the voice service or the telephone provider.
- The agents connect to the customer's billing software and dial the claims that need work. Up to 100 agents can run at the same time for one client.
- After each call the agent writes a summary with the next steps into the customer's workflow.
- The founders brought billing operations experience (Suhail Parry), an earlier AI medical scribe company (David Day) and clinical AI research at Mayo Clinic (Yash Raj).

Sources: [1, 4]

## Who tried this before LunaBill, and why did they fail?

Three companies in our database of failed startups tried to automate hospital administration or insurance work before voice models could hold a phone call. The values below are from that database and are not independently verified.

The failed startups come from our database of dead startups, retold here in our own words; Its entries are AI-assisted summaries of public sources, so they carry no evidence grade.

### Olive AI (2012 to 2023; funding US$852M; cause: Unit Economics)

- Promise: An AI workforce of software bots for hospitals: prior authorizations, claim denials, eligibility checks and billing reconciliation.
- Why it died: The bots were brittle robotic process automation that broke when hospitals changed software or forms. Each deployment needed custom integration work, so costs grew with every customer.
- Record: 

### Opkit (2021 to 2024; funding US$1.5M; cause: Ran Out of Cash)

- Promise: Automated health insurance verification and processing for healthcare providers.
- Why it died: Healthcare organizations adopted new tools slowly, integration with existing systems was hard, Epic offered bundled services, and the company ran out of runway before it secured long contracts.
- Record: 

### Medisas (2013 to 2016; funding US$2M; cause: Competition)

- Promise: A cloud platform for hospital patient data that replaced paper records.
- Why it died: Long hospital sales cycles, compliance costs and the dominance of Epic and Cerner kept it from growing its customer base fast enough to pay its way.
- Record: 

## What changed

Olive AI and Opkit tried to automate hospital billing work with software bots that clicked through screens and web portals. Phone calls stayed with people, because a bot could not talk to an insurer's representative. Olive's bots also broke whenever a hospital changed a form, and each customer needed custom work.

Voice models changed the phone part. A model fine-tuned on insurance calls can handle a claim-status call, an appeal follow-up or a reprocessing request, which was the slowest part of a biller's day. LunaBill connects to the billing software that teams already use instead of replacing it, which was the problem Medisas faced against Epic and Cerner.

## Can one person copy LunaBill now?

**Verdict: Not by one person without healthcare billing experience.**

The product needs knowledge of claim denials and insurer phone menus, a large set of real call transcripts to train on, patient-privacy agreements before a hospital will share data, and long enterprise sales. The founders brought billing operations experience and contacts. Infinitus already sells AI agents for phone calls in healthcare.

1. Find a phone task that a whole department repeats all day with the same script, and measure its cost per call.
2. Start where you have worked: the founder's own billing background gave the vocabulary and the first contacts.
3. Offer a pilot on the customer's real backlog, so the result is in money they can count.

**What is still open.** The same pattern fits other industries where staff wait on hold with large companies: insurance claims for car repair shops, freight and customs inquiries, utility billing disputes for property managers. Each needs its own call transcripts and its own rules on automated calls.

## What could go wrong for LunaBill?

- **What is not verified.** The headline, US$428,000 of live annual recurring revenue out of US$764,000 contracted, rests on one founder statement on the company's Y Combinator page. The text has no date; this page dates it November 2025 because it first appears in the Internet Archive copy of 23 November 2025 and is absent from the copy of 19 October 2025. No payment processor profile, filing, investor statement or press report confirms it, and no later revenue figure was found. The founders said the contracted rest would go live by January 2026; no source shows whether it did. A reader could look for a press report on a later funding round or a second interview with revenue figures. The company's own website answers automated readers with a bot check, so no company page was used.
- **Do the AI callers say they are AI?.** No page we could read says whether LunaBill's agents tell the insurer's representative that they are AI. The launch post of November 2025 says the agents 'sound completely natural' and invited readers to listen to a real call and 'see if you can guess who's human'. For comparison, the sample calls on the home page of Infinitus, a company in the same field, open with 'I am an AI agent on a recorded line'. The calls go to insurers, not to patients.
- **Insurers' rules on automated calls.** Insurers can set their own rules for automated callers on their provider lines. This page did not find a source that lists those rules: the US Federal Communications Commission and Department of Health and Human Services sites refused this pipeline's automated reader on 5 November 2026, and the company site could not be read, so neither a regulator document nor the company's own policy was read. If insurers block or restrict AI callers, the product loses value.
- **HIPAA and patient data.** The calls discuss patient claims, so they carry protected health information under HIPAA, the US health privacy law. Hospitals usually require a business associate agreement and a security review from such a vendor. The Y Combinator page and the launch post do not describe LunaBill's compliance status, so it is not verified here.
- **Competition.** Infinitus sells AI agents that make and take phone calls in healthcare, as the sample calls on its own site show: calls to provider offices and calls from health plan members. The founders of LunaBill came from billing companies, and large billing firms are both customers and possible competitors if they build their own agents. The sources read do not give market shares.
- **Dependence on model and voice providers.** The sources do not say which language model, speech recognition or synthetic voice LunaBill uses, only that the agents are fine-tuned on more than 1.2 million call transcripts. If the company rents these from outside providers, their prices and terms set its costs. The transcript data set is the part the company owns.
- **Healthcare product, not a medical claim.** LunaBill works on billing and insurance administration. The sources read make no diagnosis or treatment claims for it, and its calls go to insurance companies, not to patients.

## Sources

1. Y Combinator company page: LunaBill, AI Voice callers for healthcare billing teams (read 5 November 2026) — https://www.ycombinator.com/companies/lunabill (grade D). Read in full (page data). Founder-written text with the ARR figures, the launch month and the call count; directory fields for team size, batch and status.
2. Internet Archive copy of the Y Combinator company page, 23 November 2025 — https://web.archive.org/web/20251123093429/https://www.ycombinator.com/companies/lunabill (grade D). Read in full. The earliest archived copy found with the US$428,000 text; also a copy of 30 November 2025 with the same text. Used to date the headline.
3. Internet Archive copies of the Y Combinator company page, 7 and 19 October 2025 — https://web.archive.org/web/20251019215038/https://www.ycombinator.com/companies/lunabill (grade B). Read in full. No long description and no revenue text; team size 3. Used to date the headline after 19 October 2025.
4. Launch YC, 11 November 2025: LunaBill - AI Voice Callers for Healthcare Billing Teams — https://www.ycombinator.com/launches/Ooq-lunabill-ai-voice-callers-for-healthcare-billing-teams (grade C). Read in full. Founders' launch post: product, 60,000+ calls, fine-tuning data, team and titles, how they met, the request for introductions.
5. Infinitus home page (read 5 November 2026) — https://www.infinitus.ai/ (grade C). Read the sample call transcripts. Used only to show a competitor and how its agents introduce themselves.
6. database of failed startups export of failed startups — https://www.our failed-startup database/ (grade C). Local export. AI-written summaries, not independently verified. Source for the failed predecessors.

Contact: aiolum@livyden.com. Run LunaBill? Send corrections or a verified figure to aiolum@livyden.com (subject: Correction: LunaBill).