Novartis CEO Vas Narasimhan: Drawn to Analytics, Grounded Expectations for AI

Yesterday, the MS/MBA program at Harvard Business School (HBS) hosted Novartis CEO Dr. Vas Narasimhan for what proved to be a captivating and wide-ranging discussion, led by Dr. Christiana Bardon (Managing Partner of MPM BioImpact) and Professor Amitabh Chandra of HBS and the Harvard Kennedy School.  Chandra co-leads the MS/MBA program together with the former head of Novartis’s early research organization, Dr. Mark Fishman, who attended yesterday’s talk, and legendary developmental biologist Doug Melton; I serve as an advisor to the program. 

Writing on LinkedIn, Bardon shared four observations about the conversation with Narasimhan:

• “His background as a scientist and physician has enabled him to engage more with the science and to make bolder bets in cutting edge fields such as radio ligand therapy and gene therapy. The most important thing is curiosity and asking questions!
• All pharma have tons of cash and are ready to do acquisitions at any time. The market environment and even the cost of capital don’t have a major impact on the pace of acquisitions.
• Discipline is key for internal and external investment and hes looking for a return on capital ~9%. He is opportunistic and would like to see all worthwhile projects move forward.
• FDA interactions continue to be productive and they have seen no delays or slowdowns.”

In addition to these highlights, I was struck by Narasimhan’s comments related to (a) AI and (b) finding the next big thing.

Grounded Expectations for AI in R&D

On the AI front, I was impressed by his somewhat muted view of the impact of AI – a view strongly aligned with the grounded reality this column has repeatedly emphasized (see here for example), and clearly distinct from the breathless visions some of the most impassioned advocates have projected.

In particular, Narasimhan cited opportunities for improving the efficiencies of some processes, particularly around the broad category of SG&A (selling, general and administrative expenses), and described additional areas where AI-mediated optimization could be helpful. 

He noted Novartis was involved in partnerships and collaborations with a number of leading AI R&D outfits, but he didn’t seem to feel that AI was on the threshold of substantively improving the efficiency of either discovering new targets or coming up with original medicines (although it can likely help a team get to an intended target or medicine somewhat faster, he noted). He also called out the inability of AI to distinguish high quality papers in the scientific literature from the many (many!) of more dubious quality.

Apparently, Narasimhan is not alone in his cautious view of AI.  He described (as I recall) a recent event or panel where other biopharma leaders (I think) were asked whether they saw AI impacting either their top line or bottom line forecasts for the next 5-10y, and none of them thought it would – although discrete opportunities for incremental impact were mentioned.

I’ve often described Narasimhan as exhibiting “refreshing candor,” especially in the very public way he’s wrestling with the promise and challenges of emerging technologies (see here, for example, and references therein); this week’s discussion offered another example of both his candor and curiosity — qualities equally refreshing to see in an industry leader.

Picking Winners

The discussion topic besides AI that caught my attention yesterday was the contrast I noticed between how he views the path to R&D success and how he views his own journey towards career success.

As he pointed out, the central challenge facing all biopharma companies is that because of patent expirations, companies constantly need to come up with new products and portfolio of products capable of propelling continued company growth.  Yet finding the next new thing, as he readily acknowledged, and as TR readers viscerally appreciate, is “really hard.”

Somewhat predictably, Narasimhan emphasized the importance of following the science; recruiting and developing exceptional talent (“it’s a people business”); and making good decisions –- in short, all the usual stuff that everyone else is also trying to do, and asserting they already do. 

Novartis, as Narasimhan describes it, seems (like most if not all other large pharmas) drawn to a deeply analytic approach to everything they do.  Bardon (as I recall) pointed out this sounded “very McKinsey.” (Perhaps not a surprise: Narasimhan spent a couple of years as a McKinsey consultant, and traditionally, Novartis has long been regarded as very much a McKinsey shop, although their current head of strategy, Ron Gal, is [like me] ex-BCG.)

The thing is, like other power law domains, pharma is largely an exception-based business, and it’s not at all clear that anyone can really “pick winners” – a point about which this column has long obsessed (consider the wild stories of pembrolizumab, bortezomib, and the GLP-1s, for starters – all discussed for TR readers here), and which even consultancies intermittently recognize (as I’ve also long discussed – see here). 

What I think you can do as a pharma leader is avoid obvious mistakes, optimize your approach to evaluation, and essentially try to reduce the costs and maximize the learning from each shot on goal, with the idea that if you can hang in long enough, and take enough reasonable shots over enough time, eventually, something will hit.  Then, as Merck’s Roger Perlmutter did with pembrolizumab, you mobilize the full resources of your organization to blow out the opportunity as best you can.

While an emphasis on contingency wasn’t a prominent feature of Narasimhan’s discussion of R&D, he poignantly emphasized the role of chance and lack of predictability when describing his own career.  He explained how he was as surprised as anyone that his trajectory took him into the CEO role, contrasting it with his experiences along the way.  At one point he said, he was supervising operations at a manufacturing plant at a desolate location in Europe, where the view from his office was a dilapidated used car lot across the street.

Of course, I imagine he was more analytic and deliberate (“agentic”) about his career than perhaps this endearing anecdote suggests, but I can also entirely imagine him staring at a used car lot and wondering whether his life had taken a wrong turn somewhere.

My suspicion is that the two domains may be more similar than he appreciates.  As I’ve discussed in the context of both Ed Catmull’s Creativity, Inc. and Cass Sunstein’s How To Become Famous, extraordinary success of both careers and programs requires not only immense talent, intense focus, remarkable persistence, and (I’d contend) an agentic mindset.  You also need exceptional good fortune — unanticipatable luck that even that best experts, armed with the most robust analytics and the latest AI, will invariably struggle to predict.