/Companies

Decimal Code

Techstars '25

Ann Arbor, US · 4 employees on LinkedIn · 2 known investors

Find your way into Decimal Code

2,666 people in our graph share verified history with the Decimal Code team — schools, employers, funds. One of them is your warm intro.

Michael Taylorunlockedknows Mike Burns · together at Pfizer (overlapped)
×6knows the team · via Pfizer
×5knows the team · via University of Michigan

Decimal Code provides revenue cycle AI solutions for healthcare organizations, leveraging machine learning to optimize medical coding, prior authorization, and billing processes. The company's platform is built on 15 million de-identified cases from 84 research hospitals and is designed for hospital finance teams and revenue cycle managers.

Also known as DecimalCode

Founders & leadership

JPJoey Panella
Joey PanellaCo-founder & CEO
JVJohn Vandervest
John VandervestCo-founder & CTO
MBMike Burns
Mike BurnsCo-founder & CMOMike Burns holds an MD and PhD and was an anesthesiologist and researcher at the University of Michigan, where he worked with the Multicenter Perioperative Outcomes Group on projects spanning basic science, clinical research, and computer science applications including natural language processing and machine learning in healthcare.

Investors · 2

Techstarslisted by TechstarsBoulder · $220K

How we know: the Techstars portfolio · open dataset · Top 3000 Accelerator Startups (no YC) 2026-08 · Not right? Tell us

Company profile

researched Sep 2026

Decimal Code is an Ann Arbor, Michigan-based healthcare software company that applies natural language processing and machine learning to medical coding and clinical documentation improvement. Its patented process creates local, site-specific models trained on a provider's own historical data, reading clinical and operative notes to extract relevant information and assign billing and diagnostic codes. The product set focuses on perioperative care — surgery and anesthesia — while the company states the system works across specialties.

Two principal use cases are described: automated medical coding, in which easy cases are batch-billed and more complex cases are routed to human coders with supporting data, and prior authorization CPT mismatch detection, which flags discrepancies between codes on authorization requests and codes on the eventual claim, as well as surgical deviations and missed procedures, before no-authorization write-offs occur. Every prediction is accompanied by a confidence score and supporting signals, which the company frames as a transparency mechanism for clinical and coding staff.

The company originated at Michigan Medicine out of roughly six years of research and development, and its models were developed using 15 million cases spanning 84 research hospitals. Its team combines clinicians, data scientists, researchers and medical billing specialists, and it has published research on automated CPT code classification from surgical notes and on machine-learning classification of anesthesiology CPT codes from EHR data.

Founding story

The company grew out of work at Michigan Medicine. Mike Burns and John Vandervest led a research consortium of 84 health systems holding 15 million procedural cases, and needed anesthesia billing codes to deliver real-time quality feedback to clinicians; because codes took weeks or months to return through traditional billing, they began predicting them directly from operative notes. After the health system's CFO saw the results, the approach was extended beyond anesthesia, leading to a process patent and 52 live models across specialties at Michigan Medicine. Joey Panella met the two in 2019 while working at the University of Michigan's Venture Center, where he supported research spinouts, and became more involved in 2021 when the team sought a commercial counterpart.

Business model

Decimal Code sells its AI coding and CDI software to healthcare provider organizations. It charges an implementation fee plus a per-case fee that varies with the complexity of the medical specialty, and positions this pricing as a discount relative to traditional medical coding before accounting for additional charge capture. The company states it aims to capture a fraction of the value it creates for customers.

Implementation fee plus per-case usage fees, priced according to the complexity of the medical specialty being coded.

Traction

As of March 2026, three health systems beyond Michigan Medicine, with one fully deployed and two actively deploying, plus 52 live models across specialties at Michigan Medicine. The company cites a Michigan Medicine revenue cycle administrator attributing $2.3M in annual recaptured revenue to previously missed arterial line placement and in-person ultrasound coding. Reported product outcomes include a 41% reduction in prior authorization denials, a 44% reduction in coding costs and a 70% reduction in coding time.

Latest developments

In a March 2026 interview, the CEO said the company was actively closing a $750,000 pre-seed round, with $525,000 committed from existing investors, intended to expand the engineering team as it scales to new sites.

▸Full profile — market position, technology, go-to-market, geography, history, risks & controversies

Market position

Positioned in revenue cycle AI, competing against vendors the company says lack procedural coverage. Its stated differentiators are procedural and perioperative expertise and the patented methodology for building custom models from a provider's own data, enabled by training data drawn from 15 million cases.

Perioperative and procedural depth in surgery and anesthesia; a patented method for building custom, site-specific models from provider data; per-prediction confidence scores intended to make model limitations explicit; and origins in an academic medical center research consortium with access to a large multi-hospital case corpus.

Technology

A patented process for building local, site-specific AI models trained on a health system's historical data, combining natural language processing with machine learning to read notes, extract salient information and apply medical codes. Each prediction carries a confidence score. The company reports that traditional machine learning outperforms large language models for the categorization task of assigning a case to one of roughly 10,000 code choices. Published company research reports anesthesiology CPT classification accuracy of up to 96.8% from EHR data.

Go-to-market

Direct enterprise sales to health systems, with demo requests as the primary website conversion path. Early pilot partners Michigan Medicine and Northwestern Medicine are also equity holders, and the company has used the procurement process learned at Northwestern to shape its sales approach at other sites. It participated in the Techstars accelerator (2025 class) through a partnership involving Northwestern Medicine.

Health systems and academic medical centers, specifically their medical coding teams, revenue cycle managers and hospital finance leadership, with particular applicability to surgical and anesthesia service lines.

Geography

Headquartered in Ann Arbor, Michigan, United States, with customers at Michigan Medicine, Northwestern Medicine and additional U.S. health systems.

History

Founded on approximately six years of research and development at Michigan Medicine, the company spun out of a team of data scientists at the academic medical center. It secured a process patent and deployed 52 live models across specialties at Michigan Medicine, then joined the Techstars 2025 class. As of March 2026 it reported three health systems outside Michigan Medicine, one deployed and two in active deployment.

Risks & controversies

The company identifies validating its models as a core difficulty, noting that national medical coding accuracy is around 83%, so discrepancies between model output and historical human coding require auditing to determine which is correct.

Compiled by commissioned research from 8 cited public sources — announcements, filings, and press listed under research sources below.

Key figures

latest reported
Anesthesiology CPT code classification accuracy (research study)Jan 202696.8%
Annual revenue recaptured at Michigan Medicine (customer-reported)Jan 2026$2.3M
Cases where the correct code is within the top 3 predictionsJan 202699%
Cases where the top predicted code is correctJan 202684%
Cases where the top predicted code is correct with confidence sufficient for autJan 202670%
Coding accuracy above 95%, share of total casesJan 202651%-92% of total cases, depending on surgical service complexity
HeadcountAug 20264
Health systems outside Michigan MedicineMar 20263 health systems
Live models deployed at Michigan MedicineMar 202652 models
Pre-seed round amount committed from existing investorsMar 2026$525K
Reduction in coding costsJan 202644%
Reduction in coding timeJan 202670%
Reduction in prior authorization denialsJan 202641%
Research hospitals contributing training dataJan 202684 hospitals
Training cases used to build the platformJan 202615,000,000 cases
Years of R&D behind the platformJan 20266 years

Company-reported or press-reported figures, each dated to when it was claimed — not independently audited.

Competitors · 2

by search overlap

Companies competing with Decimal Code for the same Google search keywords, organic and paid, via search-intersection analysis.

Timeline · 4

launches, deals, and filings
Mar 2026
Techstars startup spotlight interview with CEO Joey Panella

Techstars published an interview covering the company's origins at Michigan Medicine, its technology, customers and fundraising status.

source ↗

Mar 2026
Deployments at three health systems outside Michigan Medicine

The company reported one health system outside Michigan Medicine officially deployed and two actively deploying.

source ↗

Mar 2026
Pre-seed round of $750K being closed

CEO Joey Panella said the company was actively closing a $750,000 pre-seed round, with $525,000 committed from current investors, to build out the engineering team as it scales to new sites.

$750K source ↗

Jan 2025
Participated in Techstars 2025 accelerator class

Decimal Code joined Techstars as part of its 2025 class, through a program involving a Northwestern partnership; the company cited Northwestern Medicine and Michigan Medicine as early pilot partners and equity holders.

source ↗

Dated company events from announcements, filings, and press; legal rows summarize public dockets and regulator releases.

▸Research sources · 8

primary sources listed

8 public sources were cited for this profile; the first-party ones are listed here.

Frequently asked questions

What does Decimal Code do?
Decimal Code builds site-specific AI models that read clinical notes and predict medical billing codes for health systems.
Who founded Decimal Code?
Decimal Code was founded by Joey Panella, John Vandervest, Mike Burns.
Who are Decimal Code's investors?
Decimal Code's investors include Chicago Early Growth Ventures (CEGV), Techstars.
Where is Decimal Code headquartered?
Decimal Code is headquartered in Ann Arbor, US.