Artificial intelligence is becoming increasingly important in the development of medications and therapies. What does this mean for organizations – for internal structures, strategic orientation, and decision-making processes?
The use of artificial intelligence (AI) is currently revolutionizing the entire life science industry. Its application in product development significantly accelerates the traditionally lengthy and costly processes of drug and therapy development, making them more precise. More and more companies are using AI-supported platforms to identify novel drug targets, predict molecular structures, automate synthesis pathways, and enable patient-specific therapies. These technologies have the potential to reduce development times from years to months while significantly lowering costs – without compromising scientific rigor. The use of AI extends well beyond product development. In Medical Affairs, AI processes are creating and managing information and evaluations. When it comes to commercializing new drugs and forms of therapy, AI is increasingly helping to analyze market data and provide control recommendations.
For decision-makers and their organizations, these new possibilities always have the same consequences, as structures, strategies, and decision-making processes have to be adapted. Members of an organization have to learn to accept new things, question routines, adapt, and be open. AI enables not only the digitization of existing processes; value creation can cut across current divisions of labor.
This dynamic can be overwhelming, overlooked, or get out of hand. But it can also set and reinforce exactly the right impulses. In addition to the technological fascination with AI, it is therefore necessary to actively shape one’s own organization in order to fully recognize, optimally utilize and scale the potentials. It is also about innovating how one’s own organization works.
From Discovery to Testing
AI application areas are extremely diverse, spanning all phases of development. In the early phase of drug development, AI supports the analysis of huge biomedical datasets to identify disease-relevant target proteins – for example, the development of a drug against idiopathic pulmonary fibrosis (IPF) by the biotech company Insilico Medicine. The company’s AI platform analyzed pathological mechanisms and subsequently generated the molecule ISM001-055, which is currently being tested in Phase IIa studies.
But for such innovations to have an impact, they also have to be accepted in an organization. For example, some companies report that management personnel are struggling to believe or really trust AI-generated results, often demanding that the results be validated using established methods. This can be sensible in individual cases or for a specific transition period. In the long term, however, an organization has to learn and find a common, transparent way to deal with the “new source of truth” AI results represent.
Later in the development process, AI can help match patients as efficiently as possible with suitable clinical trials. This can involve, among other things, analyzing electronic health records, genomic data, and participation criteria. AI thus shortens recruitment times, lowers costs, and increases the diversity of the participants. After all, it is often difficult to find enough suitable participants, or the dropout rate is high. AI systems can also identify potential study participants who might respond better to certain therapies based on specific biomarkers. Such stratified studies, with patients divided up into subgroups, are an important component in the development of personalized medicine.
At the latest at this point in time, companies like Insilico Medicine have to ask themselves whether the AI platform is just a solution for better product development – or has a medical product for diagnostics also been developed. Can and should the company change its value creation and also focus on diagnostics?
Virtual Cells and Personalized Cancer Therapies
Medical advances enabled by AI were a key topic at the SXSW technology conference in Austin, Texas in March 2025. Dr. Priscilla Chan, co-founder and director of the Chan Zuckerberg Initiative (CZI), presented the organization’s work on virtual cells. With the help of AI, the CZI team is aiming to create a kind of digital twin of the human cell within the next few years. “Currently, the development process is cumbersome, takes years, and consumes billions of dollars,” Chan said on the conference stage. “Side effects are often discovered late in the process and for rare disease research, the numbers often don’t add up because there are too few cases.” While there is still a long way to go to develop virtual cells with the help of AI, the benefits of such technology can scarcely be overestimated. “Medical progress tends to happen in big leaps,” Chan said. “In one fell swoop, microscopes, X-rays, or genomic analysis have allowed us to understand our health so much better.” CZI represents a new kind of cooperation between the academic-medical world and the tech industry. AI expertise from Mark Zuckerberg’s Meta is flowing into research on virtual cells, while CZI’s financial resources enable a gigantic computer network of 1,024 of the currently most powerful Nvidia processors to be funded – an investment hardly feasible for universities.
This raises the question of how such gigantic technological upheavals, and new stakeholders like CZI, are changing the organizations they impact. To better answer this question, Metaplan conducted in-depth interviews with 27 executives from the pharmaceutical and biotech industries. An important insight from these conversations was that the most important long-term cost driver could be the generation, processing, or purchase of data. Similar to printers, where ink and toner are the actual cost factors rather than the devices, costs in the AI field are more likely to arise from the acquisition, maintenance, and management of data.
For organizations this means they also have to keep an eye on the cost and complexity of such a fundamental implementation of this new technology. Although the fear of missing out on AI is currently particularly strong, an early focus on the return on investment in the experimentation phase will narrow down the use cases too much.
A clear answer emerged from the 27 interviews on the question of where AI should be positioned as an innovation topic in the organization. In multiple cases, a cross-functional steering committee ensuring that AI initiatives comply with organizational goals, regulations, and ethical considerations has proved to be successful. Many organizations that initially relied on a single AI-responsible person switched after a while to such a steering committee model.
The Opposite of Moore – Eroom
During an SXSW panel discussion titled “AI Big Bets in Health Care”, Karen De Salvo, Google’s Chief Health Officer, and Diogo Rau, Digital Chief and Head of Research and Development at Eli Lilly and Company, also agreed that AI is enriching the life science industry and will do so even more in the future. Whether overworked staff, expensive and hard-to-access therapies, or the ever-increasing amount of medical data are the issue, AI offers excellent possible solutions for all these completely different problems. “About a third of all data generated annually comes from the health sector,” De Salvo said. “But with AI it’s possible to master this information overload. After all, AI is much better than us humans at analyzing large amounts of data, recognizing patterns, and thus also developing and testing new medical hypotheses.”
Diogo Rau, who came to pharmaceutical company Eli Lilly from tech giant Apple, focused on what the two industries can learn from each other. Whereas in the computer industry Moore’s Law applies, which says that chip performance roughly doubles every two years, the opposite is true in the world of medicine. That the development of a new drug has become exponentially more expensive and time-consuming since the 1950s is referred to as Eroom’s Law – Moore backwards. AI has the potential to change this, not least because the technology is accelerating stronger collaboration between the digital and medical sectors. The tech industry moves at an entirely different speed than all other industries, but nowhere is this more evident than in the pharmaceutical industry. “Some of my meetings take place quarterly and many projects have a time horizon of ten years,” says Rau, who switched from Apple to the pharmaceutical industry in 2021. “In the tech industry, meetings are at the most on a weekly basis – and ten years is science fiction.”
This striking example illustrates that AI is not just another technological innovation, like perhaps the switch from floppy disks to USB drives or from cable modems to Wi-Fi. In nearly all cases, artificial intelligence will really transform tasks and work distribution in companies – and change corporate culture. Organizations that are already preparing for this change will clearly have an advantage. The same applies to leaders who manage to alleviate their teams’ fears or concerns about AI and honestly discuss the opportunities and benefits along with the costs and risks. Ideally, the transformation toward an organization that uses AI sensibly and profitably will then take place not only on the “show side” (i.e., externally) but also be defined by official rules (i.e. the “formal side”). And on the “informal side” it will also be lived out by the employees.
