Warburg Pincus trains a model on decades of its memos to argue against each new deal

The scenario rule and the memo model check different things. The model learns from Warburg's own memos going back decades ^[6], and before the firm made an AI view mandatory, a memo did not have to contain one ^[12]. So most of what the model has learned comes from deals written up without an AI view. On disruption from automation, the forward-looking scenario has to do most of the checking ^[12]. "What we do is on every deal we look at today, the investment memo is fed in, and it tells us... what's the case against? Why shouldn't you do this?" Mahadevia said ^[7]. "It won't replace our judgment, but it provides another voice... in the room to help us think about things based on the past and what's happened," he said ^[8].

For sellers the rule could go three ways. It could move price, so that a BPO or IT services company with no answer to automation gets a lower bid or none. It could turn into a paragraph every memo carries and no vote turns on. Or the case-against model could become the dissent partners rely on, and weak deals would die before price comes up. Mahadevia did not describe how sellers or portfolio companies have responded, and the account does not name a deal the rule stopped. The thesis must also say how AI can be deployed to drive operational alpha ^[1]. A seller's defence against AI and its pitch for AI are therefore graded in the same memo. I think the rule will show up in price in the two sectors he singled out ^[5], because he framed it as a condition of investing at all. "We cannot invest in any deal unless there is a view on how technology and AI will have an impact on that business," he said ^[3].

The counter-case comes from Mahadevia's own view of India. He acknowledged the worry that AI could erode the country's demographic dividend by disrupting IT services and BPO, then argued the growth story has "multiple decades" of runway ^[10]. He also said that despite "eye-popping" valuations, relatively few funds in India can do value-creation work at scale, especially compared with the US ^[11]. A buyer that competes on operating skill can use the operational-alpha half of the thesis to justify paying up. In that case the disruption scenario costs a seller nothing. If Warburg keeps closing Indian IT services and BPO deals at the valuations other funds pay, the scenario is not moving price and I am wrong.