VC firm General Catalyst wants to buy a health system

While it’s widely accepted that the pharma industry is innovative in R&D, it is also true that it can be slow at embracing technological revolutions. Many people have criticized pharma companies for being slow to adopt AI. Indeed, some CEOs I talk to are concerned about too widely adopting AI, citing fears of unknown threats.

But as the CEO of Sanofi, I don’t believe those challenges should guide our thinking or adoption of AI in the pharma business, as AI has the potential to improve and reinvent the way our business operates. AI impacts the way we exchange information by connecting different business units and functions that may have been operating independently in real time. The exchange of data in real time greatly accelerates and enhances the business operations.

advertisement

But we have to be thoughtful about how we use AI. While some see AI adoption as a way to improve efficiency — for example, through the acceleration and automation of repetitive tasks — I think its real promise lies in insights and better decision intelligence, which will translate into better medicines, quicker, for the right patients, ultimately improving people’s lives.

Discovering innovative medicines is increasingly challenging, and the bar for differentiation and safety and efficacy is getting higher. “Expert AI” is about giving scientists the opportunity to benefit from massive computing power, machine learning, and trained algorithms for expanding the druggable universe. AI-enabled screening can sift through billions of possible molecules can allow R&D teams to shorten the search to find disease drivers and potential drug candidates. We can also be broader in the diseases we target for low incremental cost.

Sign up for STAT Health Tech

How tech is transforming health care and the life sciences

The development of expert AI for the discovery of new medicines will concern a relatively limited portion of the pharma workforce: those with highly specialized skills and knowledge at the intersection of biology, chemistry, data science and in collaboration with tech and startup companies.

advertisement

But expert AI is not the only approach that pharma should take. The promise of using AI exists across the value chain of the industry as it can deliver insights and ultimately better outcomes, end to end across the enterprise. That’s the importance of fostering the use of “snackable AI” inside organizations.

“Snackable AI” is about applications that everyone across an organization can use on a daily basis. Apps can capture and aggregate a 360-degree view from the whole company, from finance to procurement and from supply to quality. Such apps can tell us what is going well and highlight where there are potentially problems. It can also give recommendations and “nudge” by suggesting next steps that can help fix an emerging problem that’s useful at that moment in time. No individual could process all that information all at once.

Moreover, senior leadership can see the data at the same time as the teams, which means it is not polished or interpreted in advance, erasing levels of hierarchy in gleaning insights. Thankfully, it also means a very different way of working, with fewer Excel sheets and PowerPoint slides! In some cases, budgeting can entail thousands of slides; AI can cut that number to several dozen. We can eliminate wasted time. If we can democratize the data, there is less time spent in meetings and more focus on getting insights fast.

Related: Generative AI Tracker: A guide to the health systems and companies driving adoption

So, while “expert AI” is about giving to specialized R&D teams bespoke tools and technologies to find breakthroughs for patients, “snackable AI” is about democratizing access to company’s data and helping the largest parts of an organization to make better everyday decisions. That will in turn lead to better allocation of resources and, ultimately, to better outcomes for patients.

The rapid progress of AI has triggered debate about the power of the technology, and risks. Of course, we have to have good governance, but we need to allow people the ability to play with AI in a curated environment. With the right guardrails in place, we can greatly enhance our decision intelligence and get breakthroughs and real added value. When people aren’t producing PowerPoint or Excel sheets or debating the numbers, they can spend more time on seeking solutions rather than simply trying to identify the problem.

The journey is as much cultural as technological. And it must start at the top. Most leaders of large organizations started their career in an analog world and must catch up to better master the basics of digital, data and artificial intelligence. Resistance to implementing AI can exist among teams for a number of reasons — fear of process disruption, wrong decisions, and job elimination, to name a few. A key for greater adoption of snackable AI relies on a leader’s ability to demonstrate how this new technology can help teams remove menial tasks and transactional work by focusing instead on ways to empower better decisions, founded on facts and less on emotions.

We need to get the base of our companies nudged into better decisions, every day, in a nonpolitical and nonconfrontational way. As we innovate with AI, the results achieved, such as more precise and better medicines and the possibility of operating in a more efficient way, will benefit many companies as the technology learns and improves. The advantage will come from those who operationalize it faster, who live it, because much of AI is about behavioral change. You can be slow and get replaced — or act fast and be brave to make AI a key driver of progress and better decision intelligence in companies.

Paul Hudson is CEO of Sanofi.