emtelligent Turns Unstructured Clinical Data Into Actionable Insights

Roughly 80% of clinical information is buried in formats that physicians can’t readily access: other clinicians’ notes, radiology reports, genomics data, discharge summaries, and, of course, the dreaded fax. These unstructured formats are beyond the grasp of most analytics systems.

For a decade now, emtelligent has been working to change that, Jack Mosey, the company’s VP of GTM Strategy and Operations, told The American Journal of Managed Care during his first visit to an annual meeting of the American Society of Clinical Oncology (ASCO).

Founded by radiologist Tim O’Connell, MD, and headquartered in Vancouver, British Columbia, emtelligent spent its first 7 years building a platform to resolve the data silo problem, as the rise of large language models (LLMs) and ubiquity of artificial intelligence (AI) contributed to the cause.¹

O’Connell, who still practices, has said that the company strives to bring clinical credibility to a space increasingly crowded with AI entrants, as well as an engineering philosophy that prioritizes data quality.^1,2

So much of what emtelligent does with data is at the core of what oncology worries about right now, Mosey explains. “If you’re able to structure that [data] and extract that information from the unstructured clinical data, you can add a lot of value in clinical research and analytics, in care delivery—and then in the payer space, we also have value in terms of risk stratification.”

emtelligent’s flagship product, the Medical Language Engine, uses LLMs and natural language processing (NLP) trained on more than 2 million expert-annotated data points to transform unstructured clinical text into structured, coded data mapped to standard ontologies including SNOMED, ICD-10, and specialty-specific classifications.² The company reports 99% benchmark accuracy and has processed more than 5 billion clinical notes across more than 10 therapeutic areas. A second core product, Document Manager, handles the intake side—splitting, identifying, and categorizing complex paper-based and digital medical records so they become immediately usable. Together, these tools underpin Clinical Workflow, an AI-assisted review interface designed for clinicians, coders, and reviewers who currently spend hours manually searching through voluminous, disorganized patient records.²

The platform serves payers, health systems, pharmaceutical and life sciences companies, and health technology and data services providers, with the latter representing a particularly important distribution channel. Mosey explained that emtelligent often functions as embedded infrastructure within larger data aggregator platforms, structuring the unstructured clinical data before it feeds downstream analytics models.

The company has achieved commercial traction with real-world evidence companies and data resellers serving pharma, Mosey said, with Optum serving as its beta customer when it came to market in 2023. Revenue is currently in the $5 million to $10 million range, he said, with approximately 50 employees.

What about oncology? Mosey highlighted 2 high-growth use cases: the first is supporting EHR-to-EDC workflows, the shorthand for the rapidly growing need in the clinical trial process, that allows automated transfer of data from the electronic health record (EHR) to electronic data capture (EDC), which is utilized by leading research organizations such as Sarah Cannon Research Institute. This process reduces burden on clinic staff and eliminates errors.

The second is improving patient identification and trial screening by surfacing granular clinical insights from records that less sophisticated methods might miss. Mosey also pointed to the need to make sense of the unstructured data being created due to the rise of AI scribe transcripts—a technology that has taken off in oncology, especially with the AI tool DeepScribe.

• First, he said, “the EHRs need that data structured and made more useful for billing purposes.
• Second, “They also need it for clinical decision support to have it in the right place, in the right format, and in a more actionable [format] for their doctors.
• A third possibility, which Mosey characterized as a “strong hypothesis,” is the potential to move the unstructured AI transcript through emtelligent’s Medical Language Engine.

“You’ll get a better quality note than if you just feed the transcript,” Mosey said. “We’re trying to figure out how to properly benchmark that—and show it in a way that will resonate with the market.”

References

1. Southwick R. Using AI to make sense of unstructured data. Chief Healthcare Executive. December 16, 2025. Accessed June 24, 2026. https://www.chiefhealthcareexecutive.com/view/using-ai-to-make-sense-of-unstructured-data
2. emtelligent unveils AI-powered workflow solution for clinical review. News release. October 29, 2024. Accessed June 24, 2206. https://emtelligent.com/press/emtelligent-unveils-ai-powered-workflow-solution-for-clinical-review/