Lightspeed Leads $37M Series B in Healthcare LLM Startup, Tennr

When a doctor refers a patient to a specialist, they want that patient to receive quality care—and quickly. But messy documentation, lengthy clinical reviews and constant back and forth slows everything down. And the document work isn’t just data entry. It’s fine-tooth reasoning over dozens of pages of clinical information at a level of complexity that has largely stumped automation systems — and kept providers battling to cross their t’s and dot their i’s in an effort to see patients faster and get paid by insurance. Now, just a few months after announcing their Series A from a16z, Tennr is expediting millions of patients through the US healthcare system with an automation platform centered around their suite of document-reading machine learning models built specifically for medical documents.

Tennr's customers are cutting pre-visit patient processing periods from weeks to hours, while simultaneously reducing insurance claim denials for providers—a crucial advantage as healthcare reimbursements shrink and costs soar. The business has now secured $37 million in funding to grow their research team and expand their sales and marketing efforts to help more providers.

Tennr's $37 million Series B round was led by Lightspeed Ventures, with participation from existing investors a16z and Foundation Capital. This brings the company's total funding to over $61 million, following a Series A raise just six months ago. In the interim, Tennr has seen hockey-stick customer growth and achieved a series of technical breakthroughs with novel techniques to scaling vision-language models.

Tennr’s approach for healthcare practices has delivered because of how the team has developed its machine learning models. The popular “AI” large language models like ChatGPT and Claude are designed to be everything for everyone, and trade-offs required to make an excellent chat-based assistant lead to sub-par performance in accuracy, cost, and speed when reasoning through nuanced medical documentation riddled with surprising edge-cases.

And in healthcare, every case starts to feel like an edge case. As one example, practices constantly either receive documents with multiple patients non-contiguously spread through dozens of pages or dozens of documents coming in all at once for a single patient. Tennr built a “Multi-Patient” model for delineating which patients appear on what pages, regardless of document size, and a patient pooling mechanism to avoid duplicate work by merging documents all denoting the same patient.

Tennr’s models are built with an obsession towards these hundreds of such “edge-cases” that completely stump the more generalized models being developed for more commercial assistant use-cases. Even when the problems they run into aren’t healthcare-specific, like reading checkboxes, the team sees no problem in starting from scratch to build something better.

“I think checkboxes are a good example of a situation where nothing on the market, paid or open-source, was even close to hitting the accuracy requirements we needed," commented Tennr Co-Founder and CEO Trey Holterman. "The forms you fill out at a clinic are mostly checkboxes and so we addressed that by building a checkbox reader. We applied novel vision techniques we’d learned about in 2022, with what had to be the world’s largest dataset of labeled checkboxes.”

Tennr has monetized their business through an automation platform that works for everyday healthcare businesses looking to automate from the moment a document is received to the time a patient is cleared with insurance and ready to be scheduled. “Our customers, spread across the US, run tight operations and process patients with excellent service. We align ourselves with these objectives that’s helping us drive growth. In the face of fixed-fee structures on the revenue side and constant inflation on the cost side—they want to be known as the best place to send patients, but they have an insurmountable amount of admin work that they have to do to get the job done. So, whether they always know it or not, it actually really matters to them that we try to be the best in the world at reading checkboxes, and drive these models forward” added Holterman.

Today, Tennr processes the documents for millions of patients across hundreds of practices, systems and groups. Darius Reid, Head of Operations at Total Medical Supply, based out of Texarkana, describes Tennr as “completely transformative to our workflow. We’re now processing new patients in a fraction of the time it used to take, which has been a game-changer.”

Over the past few months —as Tennr has built out processes for automated intake, clinical audits and reviews, requests for more information, prior authorization requests, and eligibility & benefit checks—Tennr is getting closer to streamlining all of medicine’s pre-visit work. In the future, they hope to help make healthcare practices communicate more like tech companies—with tight, structured API responses, automated messaging, and clean data transfers. And yes, it can all still start with a fax. Having an all-in-one healthcare document processing and workflow automation platform means practices prevent more errors upfront before a patient visit even happens. They waste less time hunting down billing errors after the fact and avoid expensive claim denials with insurance. And crucially, a tiny fraction of the staff members are required to be involved in manual document processing—improving efficiency and preventing burnout.

With this funding round, like their customers, the company is growing their research and engineering teams, and plans to expand into new specialty practices still manually working their fax queues. In the next year, Tennr anticipates expediting the movements of over 10% of all Americans’ being referred throughout the US healthcare system.

"It's clear that Tennr's product is meeting a significant market need across the healthcare industry," Alex Kayyal, partner at Lightspeed commented. "Their workflow automation platform drives significant ROI for customers while improving the patient experience dramatically."