By Rotem Gura Sadovsky, Senior Director, Head of Data Strategy at Corundum Systems Biology
The dramatic advancements in artificial intelligence in recent years have made the integration of AI tools a strategic necessity for researchers and innovators looking to attract investment.
From our vantage point as investors savvy in computational biology, it has become clear that modern AI already contributes value to biotech companies in two different ways. The first is orchestration of existing capabilities: AI agents and language models that automate operations and connect with the tools scientists already use. The second is developing new capabilities: self-supervised models trained directly on biological or physiological data that can reveal patterns or generate designs beyond what researchers identify on their own.
Orchestration and Automation: AI as the Connective Tissue Within Biotech Organizations
Orchestration and automation improve how companies function and execute. It is now a baseline expectation that for a wide range of administrative tasks, like triaging emails, summarizing meetings, and tracking project execution. However, AI will play a greater role in addressing organizational cohesion, such as through maintaining organizational knowledge.
Traditional internal wikis or documentation hubs often fall out of date or lack consistent adoption. AI systems can transform this process, connecting channels where work actually happens—Slack, email, documents, and lab notebooks—and building an automatically updated, accessible knowledge base. Instead of navigating webpages, individuals simply state what information they want to add or retrieve, and the system handles the structure.
One particularly impactful orchestration task is explaining experimental results through mining unstructured notes. These notes often come in free-text form jotted down by scientists to capture experimental details that don’t fit within predefined structured parameters. As startups scale, it is unrealistic to manually inspect these unstructured observations and the valuable context is lost. AI can review electronic lab notebooks, query unstructured notes and help explain unexpected and seemingly conflicting results.
New Capabilities: Self-Supervised Models Operating on Biological Data
The second category of AI value is still maturing, but nonetheless transformative. It includes models trained not on text but on biological or physiological data, enabling predictions and generative outputs that humans could not produce unaided.
Companies like Isomorphic Labs and Evolutionary Scale use protein language models that generate novel enzyme sequences matching desired biological functions. Other companies and academic groups are developing genome language models. The Arc Institute’s Evo models are trained on bacterial genomes and evolutionary trees to predict effects of mutations and generate new antimicrobial peptides.
Continuous physiological monitoring is another promising domain. Self-supervised models trained on time-series data—like Pheno.AI’s, built from continuous glucose monitoring data—can predict future trajectories in a personalized way. As continuous biosensors are developed for additional biomarkers, new AI models could enable a richer understanding of individual physiology. For example, tracking cortisol and melatonin cycles throughout the day could help in sleep and stress optimization, while monitoring progesterone and estrogen could help improve fertility treatments.
While modern AI carries big promise for the future of biotech, it already provides value now. Startups can set themselves up for success in leveraging transformative AI in the future by adopting the imperfect tools of the present, benefitting both from the value they bring and from the adaptation of their own organizational cultures to be more AI-compatible.
About Rotem Gura Sadovsky
Rotem is a data science leader and the head of data strategy at Corundum Systems Biology. Previously, Rotem operated in computational biology and product management functions in early-stage biotech startups. As one of the first data scientists at Finch Therapeutics, he played a leading role in developing Live Bacterial Products for clinical use. Rotem’s domain expertise spans biomolecular omics technologies, clinical data from human cohorts and biomarker discovery. He holds a PhD in computational and systems biology from MIT.
