The regulatory filing sits at the heart of every biotech founder's nightmare—thick binders of compliance documentation, armies of specialized consultants, and timelines measured in years, not quarters. It's the sort of high-stakes paperwork that can make or break a company before a single patient ever sees its drug.
Panacea, a San Francisco startup that graduated from Y Combinator's Spring 2026 batch, believes there's a shortcut. Or at least a faster lane.
The company's pitch is straightforward, maybe deceptively so: use AI to handle the bulk of the tedious regulatory grunt work, then hand off the final polish—the judgment calls, the strategic finesse—to former FDA insiders who know exactly what reviewers want to see. The result, they claim, is regulatory submissions delivered at a fraction of the usual cost and timeline.
Whether FDA reviewers will warm to AI-drafted dossiers as quickly as the companies filing them is another matter entirely.
Selling Outcomes, Not Hours
Panacea positions itself as "AI-Native FDA Regulatory Services," a label that tries to capture two shifts happening simultaneously in the sector. First, the FDA itself has been ramping up its own use of AI—last May, the agency announced expanded AI capabilities and wrapped up a major data platform consolidation. A year before that, it rolled out "Elsa," an internal generative AI tool meant to help scientific reviewers and field investigators work faster.
Second, there's growing impatience with the traditional consulting model: specialists billing $400 an hour to comb through guidance documents and draft sections of an Investigational New Drug application. Panacea wants to flip that. Instead of hourly fees, it charges fixed prices tied to milestones. Clients pay when deliverables ship, not when consultants log time.
It's a bet that founders will trade some control for speed—and that Panacea's AI can actually shoulder the load.
The service spans the major regulatory pathways: drug and biologic submissions (INDs, NDAs, BLAs, orphan drug designations), medical device filings (510(k), De Novo, PMA), and various breakthrough designations. On paper, it looks comprehensive. In practice, the company is still building its case history.
The 12x Speed Claim
According to a testimonial featured on Panacea's website—attributed to Harun Sinha, identified as a principal consultant—the platform "delivered client work 12x faster" with "error-free outputs so far." Twelve times faster is the kind of number that turns heads. It's also the kind of number that invites scrutiny.
The testimonial appears on the company's own website and hasn't been independently verified. Panacea hasn't yet disclosed named clients or detailed which submissions have cleared FDA review under its watch. That may simply be a function of timing; regulatory approvals take months, sometimes years, and the company only recently completed Y Combinator. Still, it's a gap worth noting.
What the startup has shared: a mid-six-figure revenue haul during its YC stint, per a LinkedIn post from co-founder Yaman Ziadeh. If accurate, that suggests early traction. Whether it translates to regulatory wins is the harder test.
An 'Anterior Mafia' Pedigree
Both founders come out of Anterior, a healthcare AI company—what one LinkedIn observer dubbed the "Anterior mafia." Haris Javed-Akhtar, CEO, holds an MD from Imperial College Medical School and ran partnerships at Anterior before launching Panacea. Yaman Ziadeh, CTO, was an AI platform engineer there and previously co-founded Baqi, a payments platform that reportedly processed over $4 million in transactions across 20-plus enterprise partnerships. He also holds a master's in AI from the University of Illinois Chicago.
It's a resume pairing that makes sense for the challenge: deep healthcare context from Javed-Akhtar, technical chops from Ziadeh. The question is whether that combination can crack a problem that has stymied countless attempts to "software-ize" regulatory consulting.
A Crowded Field, Different Approaches
Panacea isn't alone in trying to bring AI to FDA submissions. The competitive landscape has gotten noisier. Parexel, one of the industry's heavyweight contract research organizations, partnered with Weave Bio in late 2025 to speed up regulatory filings; by April 2026, they'd expanded the collaboration to cover full NDA workflows. A handful of startups—Dossian, Complizen, Cruxi Regulatory, Regulos—are chasing similar territory, though most are selling software tools rather than bundled services.
Panacea's approach is closer to an outcomes business than a SaaS play. You're not buying a seat license to draft your own 510(k); you're outsourcing the whole thing and paying only if it ships. That shifts risk onto Panacea's shoulders, which could be a selling point for cash-strapped biotechs—or a red flag if timelines slip and deliverables stall.
The old guard—firms like NAMSA and RQM+—still command the high end of the market, with deep benches of ex-FDA reviewers and case libraries stretching back decades. Panacea is wagering that a lean, AI-first team can beat them on speed and price, especially for startups that can't afford a six-figure retainer upfront.
It's a wager that depends on one big assumption: that the AI can genuinely handle the bulk of the work without introducing errors that a human reviewer would catch. So far, the company says it's seen "error-free outputs." That's encouraging, if preliminary. Regulatory filings are unforgiving; a single misstep in data formatting or citation can trigger months of back-and-forth with the agency.
What Happens When the FDA Reads AI-Drafted Submissions?
Perhaps the bigger unknown is how FDA reviewers will respond. The agency has been cautiously pro-AI in its own operations, but that doesn't necessarily mean it's ready to embrace wholesale automation on the industry side. Reviewers are trained to spot inconsistencies, and they've seen every trick in the consultant playbook. An AI-generated submission might breeze through on technical merits—or it might raise flags precisely because it lacks the idiosyncrasies and judgment calls that experienced consultants bake in.
Panacea's Demo Day pitch apparently struck a chord with early customers, at least enough to generate mid-six-figures in contracts during a three-month accelerator program. What comes next is harder: proving the model works not just in theory, but in the messy reality of FDA review cycles, amendment rounds, and eventual approvals.
For now, the startup is making a familiar bet in the AI era—that machines can handle the repetitive, the formulaic, the exhausting, and that humans can focus on what machines still can't quite grasp: nuance, strategy, and knowing when to push back on a reviewer's question. Whether that division of labor holds up under the weight of actual submissions will determine if Panacea is a genuine shortcut or just another hopeful pitch in a field littered with them.
