2 min read
There's a conversation that happens in pathology departments across the country, usually during a budget cycle or a system migration. Someone asks: "Which speech engine should we use?" The debate goes on for a few meetings. People have opinions. IT weighs in. Eventually a decision gets made.
And then, months or even years later, the pathologists are still spending too much time on reports. Why didn't the workflow become more efficient? Wasn't speech supposed to streamline the process?
The speech engine wasn't the problem.
What a Speech Engine Actually Does
Modern enterprise speech recognition (such as Dragon Medical One, Solventum Fluency, or Microsoft Dragon Copilot) does one thing exceptionally well: it converts spoken words into text with high accuracy and very low latency. That's genuinely useful. It's also, at this point, a commodity.
Each major platform gets pathology vocabulary right. Each adapts to individual voices over time. Each delivers sub-second recognition in a quiet sign-out room. If you put three pathologists on three different engines and asked them to rate accuracy, you'd get three roughly equivalent answers.
Comparing speech engines today is a little like comparing bottled water brands. Preference is largely personal, and the choice rarely determines where pathologists actually lose time.
Where the Time Actually Goes
After a pathologist finishes dictating, they still have to navigate to the next fields, confirm required structured elements, check that laterality is consistent across sections, verify that CAP synoptic fields are complete, and work through whatever sign-out steps their AP system requires before the report can be finalized.
In a complex subspecialty case (such as a resection with multiple specimens, each requiring its own CAP cancer protocol) the actual dictation might represent less than half of the time spent on that report. The rest is navigation, structure, validation, and finalization.
Voicebrook addresses those workflow interruptions, with clients regularly reporting a time savings of at least an hour per day, per pathologist. That's time that goes back to diagnosis, not documentation.
“Voicebrook saves me at least two hours a day," says Dr. Elgida Volpicelli, a pathologist at Stamford Health. "It's not only efficient, it's more accurate and allows you to have control over your reports, beginning to end.”
This is the part of pathology reporting that speech vendors don't address. Not because it's a hard problem to notice — but because speech vendors are not pathology vendors. They build platforms for clinical documentation broadly: ambulatory, inpatient, primary care, etc. Pathology is one workflow among many, and pathology-specific friction isn't visible at that altitude.
The Layer That Matters
Voicebrook's pathology reporting solutions are not a speech engine. Our pathology-specific reporting workflow platform, PRO (Pathology Reporting Optimized) sits above whatever speech engine a department is already using.
That distinction matters in practice. For example, when a pathologist opens a case in Epic Beaker, PRO pulls case context from Beaker, such as accession number, specimen list, cassette identifiers, and prior reports. It organizes the reporting workflow around it. Templates are pre-structured for the case type. Required fields are built in. Validation logic runs in the background: laterality checks, staging completeness, CAP protocol compliance. Voice commands don't just insert text — they navigate template fields, drive Beaker actions, and finalize cases, hands-free, from open to sign-out.
The speech engine provides the recognition layer. Voicebrook provides everything else.
This also means organizations already running a speech engine don't have to abandon that investment. Already using Dragon Medical One? PRO uses the existing DMO licenses. The engine pathologists know continues to run. What changes is what happens after the words are captured.
The Right Question
The question worth asking isn't "which speech engine should we use?" It's "how much of our pathologists' time is spent on reporting work that isn't diagnosis?"
The gains from getting that time back — recovered diagnostic capacity, improved report accuracy, better downstream data for registries and analytics — dwarf anything a speech engine upgrade can deliver on its own.
“With Voicebrook, we get time savings, accuracy, and improved patient safety," says Dr. Raed Sulaiman, Senior Pathologist and Medical Director at Physicians Laboratory. "The flexibility and freedom saves us time, and that time is important."
Speech-to-text is solved. The reporting workflow is where the work is.
Uncover the hidden inefficiencies in your workflow and see how Voicebrook can help.
