Working takes on AI inside Indian pharma.
Editorial essays from the AI for Pharma organising team. Published as the conversation warrants, not on a content-marketing schedule. Each piece tries to say one specific thing that is true about the second half of pharma’s AI adoption in India, with the confidence and the caveats that a peer would want.
How to measure ROI on AI in pharma sales and marketing.
AI ROI in pharma commercial is hard to prove for a real reason: the thing you are trying to move, prescriptions, arrives months later and through someone else's data. Here is the unit that holds up in a board review, and the discipline that earns it.
14 June 20268-minute readRead the pieceWhere to start when your commercial data is not ready for AI.
The most common reason a pharma AI project stalls is not the model. It is that the sales, prescription, and market numbers never agreed in the first place. Here is how to start anyway, without a two-year data programme.
11 June 20268-minute readRead the pieceBuild or buy: how to decide on AI for a pharma commercial team.
Every commercial leader in Indian pharma is being told to build, and being sold something to buy, often in the same week. The decision is not a coin toss. It turns on two questions, and once you answer them the rest is procurement.
8 June 20268-minute readRead the pieceWhy Indian pharma's AI adoption is mid-way through, and what changes in 2026.
A note on what 'mid-way through' actually means inside Indian pharma boardrooms, and three forces that make the next twelve months look different from the last twelve.
25 May 20269-minute readRead the pieceThe org chart that should sit behind your AI roadmap.
Most AI roadmaps inside Indian pharma underperform because they target the wrong layer of the org. Here is the layer model that does work, and the data that lives at each layer.
25 May 202611-minute readRead the pieceThe MR-force productivity question, after generative AI.
The most expensive resource an Indian pharma company deploys is the Medical Representative. The most pitched AI workload is one that promises to make MRs more productive. Here is what that pitch usually misses, and what the version that actually ships looks like.
25 May 202612-minute readRead the piece