Just published a contrarian take on so-called “AI losers” that I think are mispriced. The mistake: assuming AI deletes demand, rather than reshaping workflows. In enterprise settings, AI usually: • automates the easy path • creates complexity (edge cases, QA, compliance) • shifts value to whoever owns the workflow That distinction matters for investors. Several incumbents are priced as if AI would erase their revenue. A more likely outcome: • demand shifts • unit economics change • incumbents sell AI back into their installed base That’s the setup. The piece looks at three names across the stack: • Teleperformance — scaled CX operations • Text S.A. — customer communication & orchestration • RWS — language, content, and IP workflows Different businesses. Same market overreaction. Why these work as a basket: • AI panic already priced in • expectations are low • outcomes don’t need to be heroic for re-ratings Two of them also pay solid dividends. This isn’t risk-free: • pricing pressure • execution risk • slower-than-expected pivots But when stocks price in unlikely catastrophes, “normal” outcomes can surprise. Full piece here: open.substack.com/pub/hydrar… Thoughtful pushback welcome. Happy to debate.