It’s probably not surprising that new pharmaceuticals go through rigorous safety research before ever being tested on humans.

What might be surprising to learn is how much of that research process is just mundane work. Think document drafting, all day long.

Faro AI is helping scientists focus on what truly matters: the actual safety research, not the paperwork that tags along.

Also, in case you missed the news! We're back with not one, not two, but six SF Tech Week events this October. 

Check out our full announcement from last week, with the list of events below.

Arek and Ethan 🦄

Before a pharma company can even touch a single patient with a new drug, it has to write a clinical trial protocol, often a 200-page document laying out how it'll keep patients safe and prove the drug works, imagining what could go wrong and then designing around it. In addition, teams have to write hundreds of other tedious operational documents, such as manuals specifying what color tube to draw blood into and when, for every single trial, from scratch. 

Faro AI solves this grunt work. Its platform offers several agentic AI solutions, including a clinical trial optimizer that analyzes a proposed trial design in real time and tells a pharma team how burdensome it'll be on patients, how much it'll cost, and how likely it is to succeed based on predicted FDA responses. As its end-to-end system automates these administrative spreadsheets and documents, a human is kept in the loop to sign off on every decision.

Check it out: faro.ai

You had an idea for an app. You started typing. Then you remembered how your last ambitious idea burned through your usage and forced you to stop.

So you only tried the safe ideas after that.

What if you could build without watching the meter?

Bolt Lite is here: $9 a month, with up to 50x more usage on Flash, free through Oct 14.

Enough room to keep going until it is right.

Go build something reckless.

This is sponsored content.

Faro AI operates on a subscription SaaS model for enterprise biopharma companies.

Scott Chetham landed in clinical trials during a PhD clinical placement at St. Andrew’s, a tertiary referral hospital in Queensland, Australia, a major hospital that handles complex cases referred from smaller clinics. Here, he gained experience in clinical trial operations as a coordinator for several trials, and later as an investigator, the person responsible for enrolling and overseeing patients in a trial. After moving into private industry, he repeatedly saw pharma teams design trials that could not realistically be run. Namely, protocols that placed too much burden on patients or hospitals. As he put it, “I got sick of constantly telling teams, by the way, this protocol you’ve designed is infeasible.”

Chetham and his co-founder, Ross Jaffe, had known each other for roughly 20 years. They met during the Sand Hill Road era, when Chetham was pitching venture firms on behalf of a company he worked for, and Jaffe was a VC. After working on and off together for years, the pair looked for something they could build together. Chetham's frustration with infeasible trial designs, together with Jaffe's investing background in biotech, pointed them toward clinical trial operations as the opportunity worth pursuing. Chetham said their roles were naturally suited to a partnership because their skills were “quite distinctly different” and “more complementary than competitive.”

Faro is leading a market that didn’t exist 18 months ago.

Pharmaceutical R&D spending exceeds $200 billion annually, but drug-candidate success rates sit below 12%, and 90% of trials fail, revealing how much value is trapped in designing effective trials.

Faro is riding on several tailwinds in both drug discovery and clinical trials optimization. 

First, a recent review of AI platforms for trial operations documented a major augmentation in the number and types of IT platforms designed to support the planning and conduct of clinical trials, confirming the rapid crowding of the category. 

Additionally, AI-driven molecular design is producing a surge in candidate molecules. In simple terms, this means using AI to figure out what a new drug molecule should look like, instead of chemists manually testing thousands of compounds by trial and error in a lab. But since 2015, roughly 75 AI-guided molecules have entered clinical trials, with about 67 still ongoing as of 2023. In other words, computational discovery is now feeding a widening pipeline into the clinical stage. Notably, no fully AI-discovered drug has yet achieved regulatory approval despite more than a decade of effort. This reinforces the idea that the constraint is often downstream in clinical development rather than in molecule generation. And as the quantity of AI-guided molecules increases, the pressure on the trial stage only intensifies.Therefore, trial optimization is the second major tailwind Faro is positioned to ride.

While the industry often frames patient recruitment as the core issue, many of those challenges stem from overly narrow population definitions. Qualitative research with trial stakeholders across Switzerland, Germany, and Canada found that "too narrow eligibility criteria" and "overoptimistic recruitment estimates" were among the most frequently cited reasons for recruitment failure, both of which are design-and-documentation problems under investigator control. This supports the thesis that the population is defined too narrowly at the protocol stage, leaving very few patients who qualify and making recruitment less central than commonly assumed.

However, the bigger blocker in all these areas is documentation. We’re talking about the slow, error-prone process of generating, reviewing, and updating the materials that keep trials moving. Per one meta-research study, across investigator-sponsored trials, protocol writing was the single largest specific cost item (a median 7.2% of total trial cost), followed by data management (5.0%), and more than a quarter of total trial costs (median 27.5%) were incurred during the planning phase alone, which is typically underfunded. 

Downstream documentation churn compounds this. For instance, a review of 21 phase III protocols found a median direct cost of $535,000 to implement a single amendment, and UK regulators processed more than 18,000 amendments in a single year, with substantial amendments taking an average of 48 days to approve. This costs sponsors an estimated $600,000 to $8 million every day. 

Faro's own platform is built around this insight. It generates protocol drafts and operational tables that once took teams weeks to write, and it also automates the workflows that follow. This includes building the Electronic Data Capture system, the software used to collect and manage patient data during a trial, and it also speeds up study startup, the lengthy process of finalizing approvals and logistics that must happen before the first patient can be enrolled. So the entire process moves faster and more efficiently. That combination positions Faro to capture the same 30% to 50% acceleration and up to 40% cost reduction that AI has shown in aggregate, but as a single purpose-built platform and not a group of solutions combined. 

Faro's focus on speeding these layers therefore targets a quantifiable, high-cost constraint, giving it a differentiated path to capture the emerging category and establish lasting leadership.