By: Chris Paquette
Founder and CEO, DeepIntent
In 2023, BioMarin launched Roctavian, a drug promising new hope for patients with severe hemophilia. Approved as a one-time gene therapy, Roctavian was designed to improve quality of life by reducing the day-to-day burden of managing the disease. However, by 2026, Roctavian had been pulled from the market after failing to drive the adoption that BioMarin forecasted. This failure happened not because of Roctavian’s safety or efficacy as a therapeutic, but because BioMarin couldn’t commercialize this breakthrough treatment well enough to sustain it as a product.
This isn’t the first example of a drug being pulled due to commercial failure. And, as AI enables and accelerates the development of treatments targeted at more precise patient groups, it won’t be the last.
AI is changing the way we think about drug development. We are already seeing how new drugs can be discovered faster and more efficiently than ever. Companies like Generate:Biomedicines and Pathos AI are securing major investment to bring new drugs to market using AI. And, while all patients may benefit from this watershed moment in R&D, patients with rare diseases stand to benefit the most.
Indeed, orphan designations (treatments for conditions affecting less than 200,000 people in the U.S.) now exceed half of all new drug approvals annually. In both 2023 and 2024 combined, more than half of CDER’s novel approvals were for rare diseases or carried orphan designation, up from 38% only a decade prior. Furthermore, according to IQVIA, over 44% of treatments being pursued in development by pharma companies (from discovery through phase I-III of clinical testing) are designed to treat rare diseases.
But for all the innovation and progress in drug discovery, there remains a major obstacle to the success of any new treatment: the cost and the complexity of getting these drugs into the hands of the patients who need them most.
In order to commercialize new treatments, pharma companies must navigate a maze of regulators, providers, insurance companies, and pharmacy benefit managers. These stakeholders all demand evidence, education and active engagement. This requires significant time and investment into the commercial infrastructure – the teams, data, and systems powering go-to-market efforts – that can amount to hundreds of millions of dollars or more.
As a result, drug commercialization has traditionally been optimized for “blockbusters” – medicines treating patient populations often so substantial that they generate enough revenue to fund substantial commercial infrastructure while also generating returns large enough to fund the next generation of R&D.
But what happens as precision medicine introduces more “long tail” therapies that address only thousands of patients?
With a growing share of therapies targeting increasingly narrow addressable patient populations, our industry is heading towards a new economic collision: drugs designed to treat 5,000 patients simply cannot support the costs of bespoke commercial infrastructure in the same way that drugs treating 5 million patients do. If we don’t solve this, we will inadvertently create an invisible gate on innovation, care and outcomes, silently affecting millions of patients across the country and world.
This misalignment suggests that modernizing the commercialization process may be one of the most important challenges facing life sciences today and over the next decade. If we want personalized medicine to deliver on its full potential, we must re-think and re-build the infrastructure that supports developing, launching and sustaining increasingly personalized therapies in market in a more efficient way.
The good news is that the same technology accelerating drug discovery today may also help us redesign the economics of commercialization.
For more than a decade, as the founder and CEO of the healthcare marketing company DeepIntent, I have watched pharma marketing teams and their partners as they have embraced a relatively new technology – programmatic digital advertising – to connect scaled health datasets with the means to act and drive commercial outcomes more efficiently. By unifying that same data with AI-driven decision-making as well as broader process automation, marketers are already using AI to drive down costs for new patient engagement and drive adoption of new treatments more efficiently.
The impact is not theoretical. On Pfizer’s Q4 2025 earnings call, executives noted that AI was helping drive increased productivity and higher marketing return on investment, with promotional activity becoming “more targeted.”
But, as powerful as this AI impact has been, marketing represents only a fraction of the work required to support successfully commercializing a drug. At a minimum, any solution attempting to redesign the commercialization process needs to bring together a dizzying array of stakeholders and functions that each play a unique and vital role in ensuring commercial success. This includes teams responsible for finding unmet patient needs, proving value through evidence, and coordinating programs to raise awareness and reduce market friction for patients and their providers.
To be clear, what we don’t need is another disconnected AI-powered point solution. What we need is a shared intelligence layer that connects the people who drive the commercialization process with the data, the insights, the capabilities, and each other more efficiently and quickly.
Said another way, we need a more intelligent and collaborative system – a system that can more quickly identify constraints on product adoption, make recommendations on what to do next, and then act when feasible — all while improving confidence. It should be precise in showing which patients and providers to prioritize for outreach, which evidence gap to close, which access barrier to address, and where exactly the next dollar and hour of human effort should go. It should be built on trusted data and able to present actionable insights while encouraging collaboration and accelerating decision making.
By better integrating the data these functions rely on and supporting them within one commercial intelligence layer, knowledge flows more easily through the organization, data becomes more actionable, and critical decisions get made faster and in more coordinated ways.
The hardest part to achieving this vision may not even be the technology – it may be up-ending the rigidity of the processes and behaviors that drive commercialization today.
But this is one area where agentic AI can be so powerful. By giving all stakeholders access to a shared AI interface, connected to a common commercial intelligence layer, we will incent the removal of technological and organizational barriers to progress, massively reduce time-to-insight, and significantly improve speed and cost-efficiency.
A new commercial intelligence layer powered by agentic AI will accelerate the flow of information, empower teams with more current data to make better, more informed and more confident decisions. These teams will drive better commercial and clinical outcomes by knowing a more precise mix of which commercial “levers” to pull – and how hard it is to pull them. These decisions will be increasingly implemented as permissioned and governed actions, underwritten by connected data and actionable insights. Faster and more accurate decisions means less wasted time and duplicated work.
But perhaps the most important effect will be that, with these efficiencies, the financial hurdle to successfully commercialize precision medicine will be substantially lowered. The minimum viable market size for new drugs will decrease, and with it, an entirely new class of targeted therapeutics will be made commercially viable. AI will not only help us discover drugs to meet the needs of patients but help us deliver them.
We are already witnessing AI’s transformative effects on the field of medicine. But it’s increasingly clear that precision medicine will require precision commercialization. AI’s effect on the commercialization process is only just beginning, but with its rich and well-structured data, pharma presents as one of the most fertile targets for AI innovation. Its application can not only improve efficiency for drug developers, but it can significantly improve the quality of life for hundreds of millions of Americans.
****
Chris Paquette is the founder and CEO of DeepIntent, a healthcare marketing company that develops AI-enabled products enabling more efficient pharmaceutical commercialization. The views expressed are his own.