From hype to reality - The future of generative AI and automation in drug discovery
Generative AI and automation are no longer just ideas on the horizon; they are now a reality. The way new medicines are discovered is evolving, and the integration of gen-AI and automation into drug discovery is increasingly seen as a foundational shift in how small-molecule drugs are designed and synthesized. Gen-AI now shapes how researchers design, make, and test potential drug candidates. These technologies are helping teams work more efficiently, but their value goes beyond productivity. Over time, they may also reduce failure rates, improve precision in targeting diseases, and help the industry respond more quickly to emerging health challenges.
Current Applications in Drug Discovery
Generative AI is increasingly being used in early-stage drug research to help scientists design new molecules. Unlike earlier methods that required extensive datasets to predict molecular behavior, generative AI can operate effectively with limited data, generating novel molecular structures tailored to specific drug targets and suggest potential drug candidates much faster. This process normally takes several years. However, we now see growing examples of successful compounds designed by AI, at a faster pace. This shift from data-intensive models to more adaptable generative AI techniques marks a significant advancement in drug discovery, allowing for more efficient exploration of chemical space and accelerating the development of new therapeutics.
This kind of progress becomes even more impactful when AI is used to design compounds specifically tailored for automation. The AI suggests a series of molecules based on their activity potential against the target disease. But rather than simply generating molecules with therapeutic potential, the AI also considers how feasible those compounds are to synthesize using robotic systems - factoring in reagent availability, reaction conditions and automation constraints from the outset.
When automation technologies work together, they form a connected cycle: AI designs the molecules, robots make and test them and the results go back into the AI system to help it improve its next suggestions. This loop allows researchers to learn and iterate quickly.
In this setup, automation doesn't replace scientists. It augments their capabilities. By making lab work more consistent, reducing errors and freeing up time for higher-level decision-making, it enables teams to focus on strategic problem-solving. At the same time, scaling up capacity accelerates project timelines and facilitates the rapid collection of relevant data, making the overall drug discovery process faster and more efficient.
Enhancing R&D Productivity
Speed enhancements aside, the combination of AI and automation in early drug discovery is changing how decisions are made throughout the process. Traditional drug discovery follows a process known as the DMTA cycle (Design, Make, Test, Analyse). While this approach has guided pharmaceutical R&D for decades, it is often resource-heavy. One of the most time-consuming steps is the "Make" phase, where chemists synthesize the molecules that AI or researchers have identified as promising. This part of the process usually involves multiple manual steps, long turnaround times, and considerable trial and error, which can delay progress and increase costs.
New technologies are helping to address this by combining AI-driven design with automation. Some research teams are beginning to implement what's known as a "closed-loop" system, where decisions about which molecules to synthesize are informed by real-time results from earlier experiments. For example, Iktos developed Makya, our generative AI platform that designs novel, synthetically accessible molecules optimized for multiple parameters, to significantly shorten the drug discovery timeline.
This approach allows for faster iteration and reduces the likelihood of pursuing ineffective paths. Unlike traditional methods that operate in isolated stages, this dynamic process enables teams to adapt based on incoming data.
The efficiency of these systems is further enhanced by incorporating real-time inventory data and supplier information. Chemists often face delays due to long shipping times for necessary reagents and materials. By integrating internal stock levels and external vendor availability into the planning process, AI-driven orchestration can prioritize the synthesis of compounds based on actual resource accessibility. This coordination minimizes downtime, accelerates project timelines, and streamlines the transition from design to laboratory execution. By aligning molecular design with logistical considerations, these integrated systems not only expedite the discovery process but also optimize resource utilization, ensuring that promising therapeutic candidates are developed more efficiently.
The potential business impact is huge. According to McKinsey & Company, generative AI could contribute between $60 billion and $110 billion annually to the pharmaceutical and medical product industries. Much of this value would come from speeding up drug development timelines and enabling smarter, more targeted strategies for identifying viable compounds.
Early implementations are showing promise. These systems are also improving success rates, helping scientists produce more viable molecules on the first try. What makes this approach especially valuable is its ability to factor in real-world constraints, like how difficult a molecule is to make, whether key ingredients are available and how much each step costs. This kind of insight is especially important when scaling drug pipelines efficiently.
Long-Term Potential and Industry Transformation
The long-term impact of generative AI and automation in drug discovery extends beyond efficiency gains. These tools enable more targeted drug development strategies, which may improve clinical outcomes by reducing late-stage failures and enhancing therapeutic specificity.
Reaching that potential, however, depends on more than just adopting the right technology. It also requires investment in infrastructure and in building internal capabilities. Pharmaceutical companies like Johnson & Johnson and Merck are now incorporating AI literacy into their R&D functions to ensure that scientists, data specialists, and engineers can work together effectively in increasingly digital labs.
Challenges still exist, but the overall direction is clear: generative AI and automation are already reshaping how drug discovery happens. What was once considered futuristic is now part of real-world workflows.
As adoption continues, the focus will likely shift from proving what these technologies can do to scaling their use across portfolios. This next phase will be defined not only by faster timelines and reduced costs but by smarter decision-making and stronger collaboration between people and machines. For organisations prepared to adapt, generative AI and automation offer a path to more agile, data-driven drug discovery.
