Building solutions for the commercial team
What AI startups are making possible in commercial pharma
What if pharmaceutical companies could conduct 12 weeks of physician research in 24 hours — and get essentially the same findings?
Or review a promotional claim in 12 seconds instead of 17 minutes, while catching issues that human reviewers sometimes miss?
Those aren’t hypothetical examples. They’re results reported by two healthcare AI entrepreneurs at MATTER’s latest AI in Commercial Pharma event.
Supreet Deshpande, co-founder and CEO of Synthio Labs, joined Ferry Tamtoro, co-founder and CEO of Revisto, for the November 5 conversation, moderated by MATTER CEO Steven Collens and produced with support from EVERSANA INTOUCH.
Physician research on demand
Conventional physician research and advisory boards can take 12–16 weeks, including recruitment, coordination and analysis. Synthio Labs uses synthetic panels — groups of AI-generated physician personas — to simulate responses to research questions on demand.
“You don’t have to now wait for 12 weeks. It happens in 24 hours and you have a lot more data points that can better inform your decision,” Deshpande said.
To test the findings, Deshpande said Synthio asks customers for discussion guides and physician panel profiles from conventional research, then runs a simulation for comparison. He reported approximately 95 percent overlap between findings from research that took 12 weeks and simulations delivered in 24 hours.
Deshpande said teams using Synthio’s tools were conducting five times as much research. A shorter research turnaround can give brand teams time to pursue follow-up questions or examine additional physician groups before deciding how to act on the findings. Teams still need to interpret the results and determine which questions require direct physician research.
Faster review with more thorough issue detection
Revisto reported completing an individual promotional claim review in 12 seconds, compared with 17 minutes for human review in one company case study. Its AI assesses promotional materials during the medical, legal and regulatory review process (MLR).
Tamtoro also reported approximately 96 percent accuracy relative to human review, measured using customers’ actual promotional materials. Revisto evaluates how comprehensively the system assesses those materials as well as the accuracy of its findings.
Perhaps even more significant than the speed improvement, Tamtoro said Revisto’s AI has identified compliance issues that experienced human reviewers missed. In comparisons with human review, he reported that the system identified approximately 5-6 percent more potential issues. The findings suggest AI could make promotional review not just dramatically faster, but more thorough.
Human reviewers remain responsible for consequential judgments. Tamtoro described how the system traces flagged issues to their sources or underlying reasons, giving reviewers a basis for evaluating its recommendations. Reviewers assess the AI’s findings and address issues before moving materials through subsequent review.
Thousands of patient and physician conversations
Synthio’s work also extends to direct engagement. Deshpande described an AI agent deployed at the American Society of Clinical Oncology (ASCO) annual meeting that could answer physicians’ questions across approximately 25–30 oncology products. In another example, he said a dermatology brand supported more than 5,000 patient conversations in 48 hours.
These applications point toward a different model for pharmaceutical engagement: one in which companies can have thousands of individualized conversations with physicians and patients, providing relevant information and answering questions at a scale that would have been impractical with human teams alone.
Deshpande emphasized that what he called clinical-grade voice technology needs to handle medical terminology correctly. A general voice system, he noted, might pronounce GLP-1 as “GLP minus 1.”
He also described controls on what the agent can answer, including responses grounded in preapproved materials and processes for directing adverse event reports to pharmacovigilance teams or off-label questions to medical information teams. People who ask to speak with a human can be connected to one. These controls are central to using conversational AI in regulated interactions.
Why specialized startups are finding a role
Deshpande said Synthio works with five of the top 10 pharmaceutical companies, an example of what he described as growing willingness to work with startups. The discussion also addressed a practical question for commercial leaders: which capabilities should pharma build internally, and where does a specialized partner bring an advantage?
Tamtoro pointed to the difficulty of building tools across organizational silos and the value of a partner that sees how similar problems are handled across pharma companies. He said insights gained from working with one company can inform work with another, distinguishing that learning from sharing customer data.
Deshpande argued that the challenge isn’t simply building an AI application. It’s maintaining and improving it as the underlying technology evolves. A company may be able to build a working solution relatively quickly, but keeping it current as models change requires sustained attention and specialized expertise. He suggested that pharma companies invest in foundational infrastructure while relying on specialized partners for applications that require continuous development.
The common thread across these examples isn’t simply that AI can make existing processes faster. It’s that dramatically reducing the time and cost of research, review and engagement can change what’s possible. Pharma companies can investigate questions they previously couldn’t afford to ask, catch issues before they slow down commercialization and engage physicians and patients in ways that weren’t practical before.
The opportunity isn’t just to do the same work more efficiently. It’s to rethink how the work gets done.
MATTER’s AI in Commercial Pharma series concludes January 27, 2027, with Andy Smith, partner at Oak HC/FT, bringing an investor’s perspective to AI in commercial pharma.