Weeks. That's how long it typically takes a regulatory team to draft a clinical study report—the dense, exacting documents that pharmaceutical companies submit to the FDA to prove their drugs work. Ritivel, a startup so new its team still fits around a kitchen table, says it can do the same job in minutes.
It's an audacious claim in an industry where caution is doctrine. But the three founders behind the San Francisco-based company, which emerged from Y Combinator's Winter 2026 cohort in January, believe they've found a wedge into one of healthcare's most document-intensive, high-stakes workflows. Their pitch: an AI platform that automates the creation of regulatory submissions without the hallucinations, opacity, or cloud security headaches that have plagued earlier attempts to bring machine learning into drug development.
Whether pharmaceutical companies—historically risk-averse and methodical—will trust a three-person team with documents that can make or break billion-dollar drugs remains an open question.
The Problem They're Solving
Every month a drug sits waiting for regulatory approval costs its maker millions in lost revenue. The bottleneck often isn't the science; it's the paperwork. Clinical study reports can run to thousands of pages. Common Technical Documents, the standard format for global drug submissions, require meticulous cross-referencing, literature searches, and narrative summaries that pull from protocols, statistical plans, and past studies.
Regulatory affairs teams are buried in this work. Ritivel's founders argue they can compress that timeline dramatically.
The platform, which went live in March 2026 across three core products, tackles different pieces of the submission puzzle. The first automates clinical study reports, pulling data from protocols and statistical analysis plans to generate complete CSRs. According to the company, it drafts non-data sections upfront, then after data lock, auto-generates biostatistics tables and figures, converting them into narrative form. "In seconds," the website promises—though perhaps with a bit of startup hyperbole.
The second product handles CTD automation, specifically the summaries in Module 2 and literature searches for Module 5, often required for generic drug applications. Ritivel claims it can search PubMed and broader sources, then draft Module 2 with proper cross-references, collapsing "days of work to minutes."
The third offering is simpler: Regulatory Search, a public tool anyone can test without creating an account. It combs through FDA guidance documents, ICH guidelines, and product-specific advisories. For paying customers, it extends to internal repositories—past submissions, standard operating procedures, archived studies.
The Technical Differentiators
In a field where one misplaced citation can trigger an FDA deficiency letter, Ritivel emphasizes two features designed to calm nervous compliance officers: deterministic outputs and word-level traceability.
Deterministic outputs mean the same input always yields the same result. No surprise changes between draft versions. No unexplained rewrites when you regenerate a document. It's a direct counter to the unpredictability that has made some early generative AI adopters wary—stories of models subtly altering language or introducing inconsistencies across iterations.
Word-level traceability is the other pillar. Every number, claim, or citation in a generated document can be traced back to its source, down to the specific word in the original file. It's the kind of granular audit trail regulators expect and quality assurance teams demand.
Deployment is entirely on-premises. No cloud storage, no external transmission, no model training on customer data. Ritivel's pitch to chief information security officers leans hard on this: the system is air-gapped ready, with 256-bit encryption, audit logs, role-based access, and single sign-on. Integration hooks into Microsoft Word for FDA-formatted output, SharePoint and Veeva for document retrieval, Outlook for automated follow-ups.
It's a conservative architecture for a conservative market.
The Team Behind It
All three founders share a background in AI research, though they arrived from different corners of the field.
Pavan Kalyan Tankala, the CEO, spent time at Microsoft Research and holds degrees in electrical engineering and AI from IIT Bombay. His academic work spans NeurIPS, ACL, and Interspeech. Most recently, he co-authored a paper on continual learning in language models presented at the ACL BabyLM Workshop in 2025.
Nirmit Arora, the CTO, also comes from Microsoft Research, where he focused on safety in agentic systems. In late 2025, he co-authored a paper examining vulnerabilities in multi-agent frameworks—relevant, one imagines, when building systems meant to navigate complex regulatory workflows.
Gunin Gupta, the COO, took a different route. Before Ritivel, he worked in management consulting at Kearney. He also studied at IIT Bombay, rounding out a founding team with shared institutional roots and complementary skill sets.
LinkedIn lists the company at somewhere between two and ten employees. Y Combinator's directory still shows three. They joined YC's Winter 2026 batch under Gustaf Alströmmer, a partner known for enterprise bets.
Timing and Regulatory Tailwinds
Ritivel's launch comes amid growing regulatory attention to AI in drug development. For years, the FDA watched from the sidelines, cautious but curious. That changed in January 2025, when the agency released draft guidance titled "Considerations for the Use of Artificial Intelligence To Support Regulatory Decision-Making for Drug and Biological Products."
A year later, in January 2026, the FDA and its European counterpart, the EMA, jointly issued ten guiding principles for good AI practice in drug development. It was a signal: regulators were moving from observation to active framework-building.
The FDA has also taken steps internally. In April 2025, the agency announced it had completed its first generative AI scientific review pilot and committed to scaling AI-assisted reviews agency-wide by mid-year. Whether that timeline held is another matter, but the intent was clear.
The market opportunity is sizable, if not explosive. Business Research Insights estimates the medical writing market—which includes regulatory documents—at $5.68 billion in 2025, growing to $11.64 billion by 2034. Regulatory medical writing accounts for roughly 42% of that, per Mordor Intelligence. The broader regulatory information management software market in the U.S. is projected to grow from around $627 million to $1.56 billion over the same period, according to Precedence Research.
Ritivel has been publishing blog posts since January 2026—analyses of clinical trial success rates, summaries of the FDA-EMA AI principles, examinations of failure modes for generative AI in life sciences. One post references the Vectara hallucination leaderboard, a nod to the accuracy obsession that defines high-stakes document generation.
A Crowded Field
Ritivel isn't alone. The regulatory AI space is thick with competitors, ranging from established enterprise giants to boutique consultancies.
Veeva Systems, the dominant player in life sciences software, announced its "Veeva AI" initiative in April 2025. The platform spans clinical, regulatory, and safety applications, with agentic AI and shortcuts baked into its Vault products. Veeva's PromoMats "Quick Check Agent," expected later that year, targets a different but adjacent workflow—marketing and legal review.
Certara joined Veeva's AI partner program in October 2025, integrating its CoAuthor generative AI writing tool with Veeva RIM. Certara claims CoAuthor cuts first-draft time by 30%. ArisGlobal has embedded generative AI into its LifeSphere platforms through its NavaX technology. Yseop joined Veeva's partner program in late 2024 to accelerate regulatory document writing.
There are also custom builds. AlphaLife Science demoed its AuroraPrime RMA tool with Veeva integration at a 2025 industry conference. SEI published a case study in December 2025 detailing an enterprise proof-of-concept for AI-assisted submissions at a large pharma company—evidence that some organizations are building in-house rather than buying.
Kivo offers regulatory submission software focused on planning and tracking, though AI doesn't feature prominently in its marketing.
Ritivel's angle is architectural specificity: fully on-premises deployment when most competitors offer cloud or hybrid models, deterministic outputs when others tolerate some variance, and word-level provenance when audit trails are often coarser. The company is also narrowing its focus to specific document types—CSRs, CTD modules, regulatory search—rather than attempting to own the entire regulatory information management workflow.
The Uncertainty Ahead
As of mid-March 2026, Ritivel has not publicly disclosed customers, pilots, or case studies. The company's Y Combinator launch post from January openly asks for introductions to pharmaceutical regulatory affairs leaders and chief information security officers—the classic early-stage signal that the team is still in active outreach mode.
Pricing isn't public. Neither is detailed funding information beyond the Y Combinator backing. A CB Insights listing suggests a convertible note round, but specifics are absent and should be treated as speculative.
The regulatory affairs community remains cautious, perhaps understandably. Anecdotal chatter on Reddit from late 2025 surfaces skepticism among medical writers about fully automated drafting tools. Some describe shuttered pilot projects. Mixed results from early generative AI deployments at companies like Novo Nordisk have tempered initial enthusiasm, even as overall market interest continues to build.
Ritivel's wager is that pharmaceutical companies will pay for speed, traceability, and control in one of the most risk-averse document workflows in existence. The platform promises to collapse timelines without sacrificing the audit trail that regulators demand—a balance easier described than delivered.
Whether that promise converts into enterprise contracts, and whether three founders can outmaneuver entrenched software giants and well-funded consultancies, will become clearer as 2026 unfolds and pilots either graduate to production deployments or quietly fade. For now, Ritivel is a bet on precision and paranoia: the idea that in a world of hallucination-prone AI, the companies with the tightest controls will win the most conservative customers.
