Plena Health (YC Spring 2026): teardown vs XMarts

In one line: Plena runs the paperwork of US specialty clinics (fax intake, referrals, patient calls, scheduling, claim appeals) with AI agents that write into the clinic's own medical records system [site] [yc]. Threat: 1, because it is the same idea (one operating layer for one type of firm) but sold to US clinics, not to our buyer [site]. A YC partner will still use it as the model answer: seven figures of contracted revenue in about 8 months, they claim [yc].

Note on the screen: the batch screen marked W5 as direct. This deep pass marks W5 adjacent, because the lens defines direct as the same kind of buyer, and clinics are not our buyer. The pattern is the same, the market is not.

1. Fact file

Item	Finding	Source
What they say	"While you see patients, we handle everything else." YC line: "the AI operating system for specialty medical practices."	[site] [yc]
Founded, team, funding	Founded 2025, based in San Francisco (one directory says Detroit). Team size not stated; two founders listed. No outside round disclosed; YC is the only listed backer.	[yc] [press]
Growth claims	"Grew 17x" and crossed seven figures of contracted ARR, across six specialties and eight states. About 80% of growth by word of mouth. All customers on annual or multi-year contracts. The launch post says 8 months; the company page says 10 months.	[yc]
Founders	Eyad Abdalla (CEO): Waterloo engineer, ex-Datadog and Shopify, then 3 years embedded in specialty clinics. Ahmed Al Mudarris (CTO): built infrastructure at AWS, earlier nuclear engineering. Both are sons of doctors. Eyad's mother's clinic was the first user. What YC likely saw: a founder who lived in the clinic plus a strong infrastructure co-founder (inference).	[yc] [press]
Customers	Logos on the home page: Huron Gastro, Hamzavi Dermatology, OSMS (orthopedics, Green Bay), OrthoSouth, GI Docs, The Respire Institute, Texas Pain Physicians.	[site]
Price	Not public. The sales job pays $225K to $325K base plus 0.20% to 0.50% equity, which suggests deals worth tens or hundreds of thousands a year per account (inference).	[jobs]
Path	No rename or pivot found. The June 2026 YC launch post is the only dated milestone.	[yc] [press]
Rivals named by others	Tennr (referral automation; $101M Series C in June 2025, roughly $605M valuation), Commure, Notable, Adonis.	[press]

2. Their proposition, reverse engineered

Part	What they do
Who buys	Independent specialty practices, groups run by management companies, and private-equity-backed groups in gastro, orthopedics, urology, eye care, cardiology and skin [jobs].
The pain	Staff spend 3 to 8 minutes per fax. The pile never empties. They say 15 to 25% of referrals were never scheduled [site]. Staff stitch work across EHRs, fax, portals, phones and email [yc].
The promise	"Staff stays on care; we run the rest." Their pitch: do not think in agents, think in workflows [site].
How it works	Base platform of reusable parts (workflow blocks, EMR connections, agent harnesses). Deep discovery with the practice, then a custom build in about one to two days, they claim [yc]. Sits on top of the existing EMR, no migration [site]. 40+ EMRs [site].
The output	Finished tasks inside the clinic's own system: fax sorted and matched to a chart, patient called the night before a procedure, cancelled slot refilled from the waitlist, down-coded claim spotted and appealed [site].
Proof style	Activity counts and small dollar figures: "25,000 patient calls a month automated", "10x faster referral processing", "$20K+ down-coded claims recovered", "under 4 weeks to go live", "$5,500 a month recovered" (all they claim) [site].
Who sells it	Demo request on the site (20 minutes) [site]. First dedicated sales hire is open now [jobs]. 80% of growth by word of mouth, they claim [yc].
Price model	Not public. Annual or multi-year contracts [yc]. Sold as outcomes ("we do the work"), not seats [press].
The story	Enemy: vendors who sell "an agent" or "an LLM" and leave staff to stitch the rest [site]. Belief: the practice needs one operating layer, and in time the EHR becomes a swappable database [yc].

3. Positioning

Category they claim: AI operating system for specialty medical practices [site] [yc].
Category a YC partner would put them in: vertical AI that replaces back office labour for one type of small firm (inference).
What they want to be known for: the layer that runs a clinic's admin work end to end, not one slice of it.
Where they sit on the map: L1 = 1, L2 = 2, L3 = 2. Buyer = sme (named customers are single and multi-site practices; headcounts not shown), business.

4. Three layers vs ours

Layer	Them (score + evidence)	Us (from the system inventory, "Marton's logs show")
L1 Company brain	1. Reusable platform and a per-practice build. "Your data trains your workflows" and "PHI never trains models" [site]. No public claim of a lasting memory that other work reads.	A vault of about 1,400 pages, 112 call transcripts and about 8,700 customer quotes, searchable by meaning.
L2 Agent orchestration	2. Several workflows (fax, calls, texts, appeals) run in the background and write to the EMR [site]. Human handover on exceptions is not described (UNVERIFIED).	About 5,600 jobs run by agents from three makers, each checked by a different AI, then automatic tests.
L3 Execution and learning	2. Does the work and measures it (calls automated, backlog cleared, dollars recovered) [site]. A 94% confidence figure shows in the demo panel [site]. Learning claim is at platform level: "each engagement improves the platform" [yc]. No public claim that measured results change the next run for the same practice.	Every failure gets a record and an automatic check (about 160). Offer feedback lines fell from 44 to 18 as 22 rules were added. Only about 15% of jobs reached the main code, and merges were unreliable in October. No outside customer.

5. Same and different

Topic	Same or different	Them	Us
Core belief	Same	One layer across the whole firm beats point tools [yc].	Same (brief): the full loop on one workflow.
Builds on top	Same	Sits on the systems a firm already uses, no migration [site].	Marton's logs show skills and the vault served over MCP; no customer system connected yet.
Who owns the result	Same	Sells outcomes, "we do the work" [press].	Target is the same: do the work and own one number (brief). Not yet proved.
Buyer	Different	US independent specialty clinics [site] [jobs].	UK SMEs and scale-ups (brief). No customer yet.
Kind of work	Different	Fax, referrals, calls, claims: rule-heavy admin with a clear unit [site].	Open: coding, calls-to-proposals, margin leaks (brief). Not picked.
Moat	Different	40+ EMR integrations, SOC 2 and HIPAA claims [site].	None in a vertical yet.
Delivery	Different	Software plus embedded engineers per practice [site] [yc].	The system plus the two founders.
Proof	Different	Dollars and counts per named practice [site].	Only our own business so far.
Speed	Different	"Under 4 weeks to go live", build in 1 to 2 days [site] [yc].	Not set yet.
Price	Unknown	Not public [site].	Not set yet.

6. Wedge overlap

Wedge	Fit	Why
W5 One layer for one type of small firm	adjacent	Same structure: one operating layer for one firm type, built once and reused [yc]. Different firm type (clinics, not accountancy or UK services). Best live proof that the structure sells.
W7 Find and fix margin leaks	adjacent	Their collections work spots down-coded claims and files appeals, with dollars recovered [site]. Same "find the leak, fix it, count it" shape, for clinic revenue.

7. Whitespace (where they stop)

Buyer gap: US clinics only. Nothing shows UK or non-healthcare buyers [site] [yc]. To own it we need a UK firm type that names the pain and signs a paid pilot.
Number gap: their numbers are volumes and small recoveries per practice ($5,500 a month, $20K+), with no before/after cost or margin per practice [site]. To own it we need one number a finance chief would sign, measured before and after.
Layer, price and speed are not clear gaps. They do and measure the work, their price is not public, and they claim under 4 weeks to live [site].

8. For the YC application

Proven today: Marton's logs show the system already runs agents from three makers that do work and are checked by a different AI, with about 160 failure records that each added an automatic check. Plena has real customers and revenue, they claim; we have none [yc].
Draft answer to "how are you different from Plena?": DRAFT. "Plena runs the admin work of US clinics. We are building the same kind of layer for a UK firm type, on a system that already checks its own work and learns from every failure." Founders say it their own way.
Future claim (needs proof): "We run one workflow for one UK firm type, live in under 4 weeks, and own one number." Needs a paid pilot with a before/after number.
What we understand that they don't: needs proof. A layer gets better for each firm when checked results change its rules, not only when a new build is reused. Plena says each engagement improves its platform [yc], so they may already know this. Do not frame it as a gap yet.
Copyability (M3): medium. Reusing a platform across clients and measuring results is a feature for them [yc]. But the healthcare moat (40+ EMR links, HIPAA and SOC 2, US payer rules) does not travel, and nothing shows them leaving healthcare. Not hard, so M3 cannot score above 3 against Plena.
Why now (M4): None found. The launch post cites rising staff costs and payer automation [yc], but gives no dated, sourced change that helps us more than them.
Do not claim: "first vertical operating layer" (Plena is one, with revenue). "We have customers or revenue." "We are faster to value than Plena" (they claim under 4 weeks; we have no number). "Company brain" as a headline.

9. Ideas worth taking

Name offers by the job, not the agent: "Sort the fax queue", "Fill the 9:40 cancellation" [site]. A buyer can picture the work.
A two-column "what vendors say vs what your staff needs" table [site]. It makes the rival's pitch look vague.
Proof tiles: one big number, one quote, the operator's job title (not name) [site].

10. Unknowns

Real price and contract size. Ask a buyer who has seen a quote; the sales job ad gives none [jobs].
Number of customers and team size. Check later job ads and a possible seed announcement.
Whether a human approves exceptions, and how errors are caught. The site shows a confidence figure only.
How much of the revenue is one-off build work and how much is a recurring fee.
The traction timeline: 8 months (launch post) or 10 months (company page) [yc].

Sources

plena.health home page [site]
YC company page [yc]
YC launch post: Plena Health, 2026 [yc]
YC jobs: Founding Account Executive [jobs]
Daily Dropout: Plena Health, the back office that runs itself [press]
On Healthcare: inside the agentic back office race [press]
ARR Club: Plena grew 17x [press]
VCBacked: Plena Health profile [press]

Plena Health

What they do

How close to us

Wedges

Who buys

Full teardown