Artificial intelligence is increasingly being used to address one of drug development’s biggest challenges: predicting whether promising therapies will succeed in humans. Olio Labs has announced research showing that its AI-powered preclinical platform can forecast potential clinical outcomes, including gastrointestinal side effects, cardiac toxicity, neuropsychiatric effects, and long-term weight changes, using advanced behavioral analysis of animal models.
Drug discovery companies have long struggled with a critical challenge: many therapies that appear promising in laboratory and animal studies fail once they reach human clinical trials. The gap between preclinical research and real-world patient outcomes has driven pharmaceutical companies to search for better prediction tools that can identify risks earlier and improve candidate selection.
AI biotechnology company Olio Labs is attempting to address this challenge with a new approach that combines computer vision, machine learning, and animal behavioral analysis to improve preclinical drug evaluation.
The company recently published research demonstrating that its AI-driven in vivo platform can predict human clinical trial outcomes across multiple categories, including gastrointestinal adverse events, cardiac toxicity, neuropsychiatric risks, and long-term weight loss.
Traditional preclinical studies often rely on limited measurements selected before experiments begin. Researchers may focus on specific biological markers, physical observations, or predefined endpoints based on historical practices. Olio Labs argues that this approach can overlook subtle behavioral changes that may contain important signals about how a drug could perform in humans.
The company’s platform takes a different approach by using AI vision models trained on expert-labeled behavioral data. Instead of measuring only a small number of predefined indicators, the system analyzes thousands of behavioral signals following drug administration.
These behavioral patterns are then processed through machine learning models trained using human clinical trial data, allowing the platform to generate predictions about potential safety issues and therapeutic outcomes.
AI Turns Animal Behavior Into Predictive Data
The central idea behind Olio Labs’ technology is that animal behavior contains more information than traditional studies typically capture.
According to the company, a single 24-hour experiment can generate predictions related to gastrointestinal adverse events, cardiac and neurological toxicity, and long-term weight-loss outcomes.
In one reported comparison, Olio Labs said its AI-driven weight-loss predictions achieved approximately 70% higher accuracy than traditional two-to-three-week studies when compared with actual human clinical weight-loss outcomes. The company also reported that the AI-based approach was approximately 20 times faster than conventional methods.
The findings highlight a broader trend in pharmaceutical research: using AI not only to discover new molecules but also to improve decision-making throughout the development pipeline.
As the cost and speed of generating drug candidates improve through technologies such as generative AI chemistry platforms and automated laboratory systems, pharmaceutical companies face a growing challenge—how to determine which candidates deserve expensive clinical investment.
“There is a surge of drug candidates entering the preclinical stage,” said David Tingley, PhD, CEO and co-founder of Olio Labs. “Drug developers need better tools to prioritize and decide which candidates should enter the clinic.”
Improving Drug Candidate Selection
The ability to predict clinical outcomes earlier could have significant implications for pharmaceutical companies. Clinical trials represent one of the most expensive and time-consuming stages of drug development, with failures often occurring after substantial investment.
AI-based prediction platforms could help researchers identify potential safety concerns earlier, prioritize stronger candidates, and reduce the number of unsuccessful programs entering human testing.
The approach also aligns with broader industry investment in AI-driven life sciences. Companies including Insilico Medicine, Recursion, Schrödinger, and Google DeepMind are developing AI systems designed to improve different stages of drug discovery, from molecule generation to protein structure prediction.
While many AI drug discovery platforms focus on identifying new compounds or therapeutic targets, Olio Labs is focusing on a different bottleneck: improving the translation between preclinical studies and human outcomes.
“We built something truly unique here,” said Tom Roseberry, PhD, CTO and co-founder of Olio Labs. “Rodents actually do predict what will happen in a human, we just weren’t measuring the right features.”
AI-Driven Combination Therapies
Beyond evaluating external drug candidates, Olio Labs is also applying its technology internally to develop combination therapies designed to improve treatment effectiveness while reducing side effects that can lead patients to discontinue therapy.
Combination therapies are becoming increasingly important across areas such as oncology, metabolic disease, and chronic conditions, where balancing efficacy and tolerability remains a major challenge.
By using AI to identify behavioral and safety patterns earlier, companies may be able to design therapies with improved patient experience before entering clinical development.
The opportunity is significant. According to McKinsey & Company, generative AI and advanced analytics could create substantial value across pharmaceutical research by improving productivity, accelerating discovery, and optimizing clinical development decisions.
Meanwhile, Gartner expects AI adoption in healthcare and life sciences to continue expanding as organizations integrate machine learning into research, operations, and decision-making workflows.
The Future of AI in Pharmaceutical Development
Olio Labs’ research represents a broader shift in biotechnology: AI is becoming a tool not only for discovering what drugs might work, but also for predicting whether they are likely to succeed.
The pharmaceutical industry still faces significant scientific and regulatory hurdles before AI prediction systems become standard practice. Clinical validation, regulatory acceptance, and reproducibility will determine how widely these technologies are adopted.
However, as drug pipelines become larger and development costs continue rising, technologies that help companies make better decisions earlier could become increasingly valuable.
For pharmaceutical researchers, investors, and biotech companies, AI-powered preclinical intelligence may represent a new layer of decision-making—one that transforms animal studies from limited experiments into richer predictive models for human health outcomes.
Market Landscape
AI-powered drug development is expanding rapidly as pharmaceutical companies seek ways to improve research efficiency and reduce clinical failure rates.
According to McKinsey & Company, generative AI could deliver significant value across the pharmaceutical value chain by improving molecule discovery, research productivity, and clinical development processes. Gartner has also identified AI and advanced analytics as key technologies influencing healthcare transformation.
The competitive landscape includes AI-native biotechnology companies such as Insilico Medicine, Recursion, and Schrödinger, as well as technology companies including Google DeepMind developing AI platforms for biological research.
As pharmaceutical pipelines expand, predictive AI tools that improve candidate selection and safety assessment are likely to become increasingly important.
Top Insights
- Olio Labs developed an AI-powered preclinical platform that predicts human clinical outcomes by analyzing detailed animal behavior patterns.
- The technology uses computer vision models trained on expert-labeled data to identify signals traditional drug studies often overlook.
- A single 24-hour experiment reportedly predicted safety risks and weight-loss outcomes faster than conventional multi-week studies.
- The platform addresses a major pharmaceutical challenge: improving confidence in which drug candidates should advance into clinical trials.
- AI-driven preclinical analysis could complement existing drug discovery platforms by improving translation between laboratory studies and human outcomes.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
