Combatting the Pharma R&D Productivity Crisis

Written By:

February 12, 2024

The Data-to-knowledge Bottleneck

Pharmaceutical companies are facing a curious juxtaposition. From 1990 to 2020, the amount of available data has exponentially increased—and this trend shows no signs of stopping. Over the same time period, company investment returns on R&D have been steadily decreasing. In 2012, average R&D returns dipped below the cost of capital. In other words, on average, investments into R&D were destructive to shareholder value. This is a huge, industry-wide problem that poses a threat to the future of research. 

The problem isn’t going away, despite pharmaceutical companies buying more data. The old belief that acquiring data will empower productive research and drive insights is not incorrect per se — it is just incomplete. The missing piece is analyzing the data. Data is not knowledge by itself — data must be transformed into knowledge. This sorting, organizing, and analyzing of ever-increasing datasets drove pharmaceutical companies to do the next logical thing: hire more data scientists. But the data keeps growing, and the scientists—now more numerous than ever — struggle to keep up, often able to only tackle the most critical of analyses. 

Now, we recognize that the problems facing pharmaceutical companies are more complex than this. Drug development efforts are now centered around rare and complex diseases that are difficult to target and treat, which doubtlessly impacts R&D returns. But while we cannot control the fact that much of the low-hanging fruit of drug discovery has been picked, we can control R&D efficiency. With a more efficient R&D process, companies can learn more faster and move toward effective treatments faster—while cutting costs. That is what’s platform accomplishes. 

Solving The Bottleneck’s platform clears the data-to-knowledge bottleneck, allowing the growing amount of data to reach its full knowledge potential. The result is faster research, enabling more insights in exponentially less time. How much less time? Researchers use our platform to shrink months of scientific work into minutes, without sacrificing quality. Our clients no longer struggle to keep up with increasing data; they use our AI to perform breakthrough research at a revolutionary speed. 

How does it work? First, the software integrates private datasets, knowledge graphs, and document stores with public datasets like, OpenTargets, and PubMed. In the generative step, we use our unique blend of gold-standard metascience and cutting-edge generative AI to produce publication-quality reports in minutes. The reports are complete with engaging, informative visuals, and are as easy to create as a query in ChatGPT or Google. Most importantly, our reports get the job done: internal stakeholders love them and they’ve been shared anywhere from internal data science expos to meetings with top therapeutic area heads. 

The platform is as versatile as it is fast. Our platform is live at some of the world’s top pharma companies, accelerating R&D research by a factor of over 100x. Use cases include target prioritization, drug repurposing, target diligence, drug safety, and novel hypothesis discovery. 

Accelerating the first step in the drug discovery process accelerates development on the whole—which is the primary goal.

Gold-standard Scientific Rigor

Speed means nothing if it is not backed by iron-clad scientific rigor, which is why we go to great lengths to ensure our platform utilizes top-quality evidence synthesis. A strong scientific emphasis is in our roots. Co-founder Spencer Hey focused his Harvard career on metascience, pioneering gold-standard, industry-leading ways to analyze biomedical data at scale. Metascience, precise and methodical, ensures all results are grounded in logic and science. With Spencer’s expertise, married metascience logic with AI’s prowess. The combination has unmatched potential for pharmaceutical development. 

“This is awesome. You’re way ahead in terms of how to think about pharma AI in an intelligent way,” said a client, a senior ML engineer at a top 3 pharmaceutical company.  

Our Method: The Opposite of Black Box AI

“Black Box AI” suggests a computer spitting out conclusions without any way to verify them, like a mad scientist accelerated to the speed of AI. Such a system would be untrustworthy, and a waste of time at best. That’s why we made sure to create the opposite: users can easily check the system’s logic to verify conclusions. Our platform’s transparent metascience empowers scientists by inviting them to take part in the process, because scientists and AI working together is more powerful than AI working alone. After all, an explainable result advances human understanding and research further than an unexplainable one. 

Scientists also maintain control over their research. Users can easily edit generated content, visuals, or text in any way they want. 

With this platform, scientists have the time and cognitive space to focus on what really matters: selecting the optimal focus for the next drug development cycle. How does this compare to the challenges you’re facing in your R&D role? Let us know at

Latest Articles


Understanding Large Perturbation Models

A brief, layperson's introduction to Large Perturbation Models (LPMs), a new tool in the drug development toolkit to simulate vast numbers of experiments digitally

Schedule a demo