The timeline for assembling a single FDA submission has become something of a running joke in biotech circles—though nobody's actually laughing. Six weeks to draft a Clinical Study Report if you're lucky. Three months for a full Common Technical Document if the stars align and your regulatory writers haven't burned out yet. One misstep, one misplaced citation, and the clock resets.
Ritivel, a young company emerging from Y Combinator's latest cohort, thinks it can compress that entire ordeal into minutes. Perhaps inevitably, the answer involves AI.
The San Francisco startup's pitch centers on auto-generating the dense, highly structured documents that drug developers must file with the FDA—Investigational New Drug applications, Biologics License Applications, the sprawling multi-module submissions known as CTDs. All of it spat out in Microsoft Word, formatted to agency specifications, with something the founders insist on calling "word-level traceability." Meaning: every claim, every data point, every assertion can be traced back to a source document in real time.
It's an audacious claim in an industry where regulatory missteps can sideline promising therapies for months or torpedo them entirely. The question—always the question with AI promises in healthcare—is whether it actually works.
Anatomy of the Platform
At its core, Ritivel offers three interlocking tools, though calling them "tools" undersells what the company's trying to accomplish. The first automates Clinical Study Reports, drafting non-data sections from protocols and Statistical Analysis Plans in seconds, then generating tables, listings, and figures once data lock happens. The second tackles the literature-heavy work of CTD Module 2 summaries, conducting searches across PubMed and the broader web (particularly for Abbreviated New Drug Applications), then drafting summaries complete with cross-references.
A third component—less flashy, perhaps more immediately useful—provides regulatory search across FDA and ICH sources, plus internal enterprise search that pulls from a company's existing document repository. There's a free trial available, no signup required, which suggests confidence or urgency or both.
All outputs conform to FDA formatting requirements and slot into existing pharma infrastructure: SharePoint for document management, Veeva for submissions, Outlook for follow-up reminders. The platform promises determinism, which in this context means same input yields same output every time. When agencies scrutinize resubmissions, that reproducibility isn't a convenience. It's survival.
The traceability piece matters more than it might sound. When an FDA reviewer questions a claim buried in Module 2.5 or challenges a statistical method in a CSR, regulatory writers need to surface the source protocol or dataset immediately—not after a frantic afternoon of detective work. Ritivel's architecture supposedly bakes that provenance into every generated paragraph.
The Paranoia Premium
Pharma companies are, understandably, paranoid. Intellectual property leaks, patient data breaches, proprietary formulations ending up in the wrong hands—these aren't abstract risks. So Ritivel's security positioning amounts to: we never see your data.
The company markets "100% local deployment." No cloud storage, no model training on customer data, air-gapped installations for companies that demand them. The security page reads like a checklist of enterprise-grade features—encryption, audit logging, role-based access, single sign-on, IP allowlisting. It's the kind of posture you need when your customers' entire business model depends on secrets.
Whether that's enough to overcome institutional skepticism about AI in high-stakes workflows remains to be seen. But at minimum, it signals that the founders understand the game they're playing.
The Builders
The founding team splits its expertise across two domains: deep AI research and actual regulatory knowledge (or at least adjacent experience). Pavan Kalyan Tankala, the CEO, spent time at Microsoft Research training language models and holds degrees in Electrical Engineering and AI from IIT Bombay. He's published at NeurIPS, ACL, Interspeech—the kind of academic pedigree that plays well in Silicon Valley pitch meetings.
Nirmit Arora, CTO, also comes from Microsoft Research, where he worked on agentic system safety. That's relevant experience given the stakes: hallucinations in a marketing email are embarrassing; hallucinations in an FDA submission can delay treatments and cost lives. Gunin Gupta completes the trio as Chief Business Officer, bringing consulting experience from Kearney and his own IIT Bombay background.
The company incorporated an Indian subsidiary on December 10, 2025 and recently posted for a Regulatory Affairs Specialist—suggesting they're actively hiring into the workflows they're trying to automate, or at least augment. According to their Y Combinator profile, the team currently numbers three, though that may not capture contractors or advisors.
A Crowded Field, Suddenly
Ritivel isn't wandering into a vacuum. Established players have been racing to bolt AI features onto their regulatory platforms over the past year and a half. Certara launched its next-generation CoAuthor platform on June 17, 2024. Elsevier introduced PharmaPendium AI on September 3, 2025 for faster submissions. Clarivate debuted an AI-powered regulatory assistant within its Cortellis suite on January 14, 2025. Weave Bio announced a partnership with Parexel on October 30, 2025 to accelerate submissions.
The difference, Ritivel would argue, is approach. Incumbents are adding AI features to existing products. Ritivel is building from scratch around AI, then integrating with existing infrastructure. Whether that distinction matters to buyers—or whether buyers even exist at scale for a three-person startup promising to reinvent regulatory workflows—is the bet.
In a January blog post, the company addressed what it called "the hidden cost of AI hallucinations in critical document creation," citing the FDA's draft guidance on AI in drug development and emphasizing provenance as the answer. The post reflected broader industry anxiety: everyone wants AI's speed, nobody wants its mistakes, and the line between those two is thinner than anyone's comfortable with.
The Demo Day Moment
Financial details remain undisclosed, which is typical for companies this early. Y Combinator's Demo Day—where batch companies pitch to investors—is scheduled for March 24, 2026, and that's usually when funding numbers surface. For now, the company's case rests on three pillars: speed that collapses weeks into minutes, traceability that satisfies regulators, and security architecture designed for an industry where paperwork mistakes can delay life-saving treatments by months or years.
Whether Ritivel can deliver on those promises at scale, with the reliability that pharma demands, will determine whether it becomes a category leader or a cautionary tale. The opportunity is real—regulatory bottlenecks cost the industry billions and patients even more. But opportunity and execution are, as founders learn, two very different things.
