Michael Gibson spent five years watching pathologists squint at slides, narrate their findings into clunky dictation software, and then watch those observations get mangled by transcription systems that had no idea what "pleomorphic nuclei" meant. By the time he left his post as VP of Engineering at Techcyte, a digital pathology AI vendor, he'd seen enough workflow friction at Mayo Clinic and other major labs to know where the real bottleneck lived: not in the image analysis, but in the tedious last step of documenting what you just saw.
So Gibson and two former colleagues—Henry Habib and Josiah Saunders—built Voquill, a voice AI scribe designed specifically for pathology reporting. The product started shipping to labs in late April, and in the four weeks leading up to its May 20 announcement, the company onboarded 25 pathologists at an undisclosed Fortune 500 reference laboratory. The core promise? Dictate your cases, let the AI route each observation into the correct field, and get back roughly 20% of the time you'd otherwise spend wrestling with report templates.
Whether that 20% figure holds up under scrutiny remains to be seen—the company claim lacks third-party validation. But the early traction suggests they may have found a wedge in a market that's been surprisingly resistant to transcription innovation beyond legacy dictation tools.
An Overlay, Not a Rip-and-Replace
Voquill doesn't ask labs to migrate off their existing systems. Instead, it sits on top of platforms like Epic Beaker, Sunquest, NovoPath, and CoPath—the lab information systems that have been entrenched in pathology workflows for years. Pathologists hit push-to-talk, narrate their findings, and Voquill's "agentic speech" engine parses microscopic descriptions, interpretations, and diagnostic comments into structured report fields in real time.
The software is designed to learn individual reporting styles over time, a feature that's meant to preserve each pathologist's clinical voice rather than flattening everything into generic templates. It's a small detail, but one that matters in a specialty where physicians pride themselves on their diagnostic language and where subtle variations in wording can carry clinical weight.
That adaptability also speaks to a broader challenge: pathologists are particular about their workflows, and any tool that forces them to standardize risks alienating the very users it's trying to help.
The Billing Code Play
Beyond transcription speed, Voquill flags CPT billing codes as pathologists dictate—a feature that the company believes could help labs capture revenue that currently slips through documentation gaps. In their Launch YC post, the founders assert that pathologists bill roughly $30 to $40 billion annually, and that "$8B+ is left on the table" due to incomplete or inaccurate coding practices, though this figure has not been independently validated.
The overall market sizing appears solid—Laboratory Economics pegged U.S. anatomic pathology revenue at $37.1 billion in 2025. But that $8 billion leakage estimate? That's a company claim without independent validation, at least for now. Still, revenue cycle optimization is a known pain point in laboratory medicine, and code-assist tools have become familiar fixtures in digital health workflows elsewhere. If Voquill can demonstrate real impact on coding accuracy, that alone might justify the software for budget-conscious lab directors.
The question, as always, is whether the AI's suggestions actually improve billing capture or simply add another layer of alerts that pathologists learn to ignore.
Early Adopters and Compliance Guardrails
Voquill says it onboarded those 25 pathologists in the four weeks leading up to its May 20 announcement, all of them at a single, unnamed Fortune 500 lab. The company hasn't disclosed the client, which is standard practice in early enterprise deals but leaves some uncertainty around how representative that initial cohort really is. Are these pathologists who were already hungry for workflow tools, or a random sample across different case volumes and subspecialties?
On the compliance side, Voquill claims HIPAA and CAP compliance, with SOC 2 certification in progress, though these claims have not been independently verified. The platform offers three deployment modes: local transcription that keeps audio on the pathologist's device, cloud routing through providers like Groq (with immediate discard after transcription), and API key configurations that let enterprises control where data flows. For labs particularly anxious about data governance—and in healthcare, that's most of them—the local option may be the deciding factor.
The codebase is open source under AGPLv3 and available on GitHub, where the team pushed desktop release v0.0.644 on May 22. That level of transparency is unusual in healthtech, where proprietary algorithms are more often the norm.
Who's Building This
Gibson's co-founders bring relevant experience, if not always from the most obvious places. Habib previously ran The Intelligent Worker, an AI education platform that claims 500,000 paid students, and spent time as a principal at Oliver Wyman managing large-scale automation projects. Saunders, like Gibson, came from Techcyte, where he led a forward-deployed team serving Mayo, Quest, and Labcorp. Two of the three founders were also founding engineers at Relay Health, a Y Combinator-backed company from a prior cohort.
That shared history in pathology AI deployments at major academic and commercial labs gives them something most healthtech founders lack: firsthand exposure to the specific workflow bottlenecks they're now trying to solve. Whether that translates into a product that pathologists actually adopt—and keep using six months later—is another question entirely.
The Market Context
Pathology labs are under pressure, and not just from the usual healthcare cost containment forces. According to the American Society for Clinical Pathology, citing Health Resources and Services Administration projections, the field faces a 7% decline in pathologist supply alongside a 16% increase in demand by 2037. That's the kind of workforce squeeze that makes efficiency tools more attractive, even to subspecialties that have historically resisted digital transformation.
At the same time, digital pathology platforms are accelerating adoption. PathAI's nationwide Labcorp rollout in February, Aiforia's Proscia integration in March—these are signals that the industry is finally moving past whole-slide imaging pilots and into production deployments. Voquill isn't competing directly with diagnostic AI or image analysis tools. It's targeting the last mile of the workflow: the moment when a pathologist has finished reviewing slides and needs to document findings.
If the early traction holds, the company may have found a wedge in a market that's been oddly resistant to transcription innovation. But four weeks of deployment at a single lab—even a large one—doesn't yet constitute proof. The real test will come when Voquill scales beyond that initial Fortune 500 client and into labs with different workflows, different LIS platforms, and pathologists who didn't volunteer to be early adopters.
For now, Gibson and his team have a product, a handful of paying users, and a thesis that pathologists are tired of spending a fifth of their day on documentation. Whether that's enough to build a durable business in one of healthcare's most particular specialties remains an open question.
