Global software equities have come under pressure recently as investors consider how newly released artificial intelligence models could replicate tasks currently performed by software vendors. With markets trying to separate potential winners from losers, volatility has risen in the sector.
Earlier in the week, strategists at Deutsche Bank described the equity landscape as a "sniper's alley," reflecting the difficulty traders face in identifying which companies will benefit and which will be disadvantaged by rapid AI progress.
Against that uncertain backdrop, Bernstein's research team has proposed a structured methodology for gauging AI disruption risk across software providers. The framework centers on two primary dimensions that together aim to capture how exposed a company's core offerings are to being replaced by AI frontier models.
The first dimension, which Bernstein strategists Richard Nguyen and Mark Moerdler label "automatability," measures whether the source of a business's value can be fully automated by AI. As the analysts explain, "Low automatability means value depends on proprietary data, complex workflows, and deep integration (difficult for generic AI models to fully replace). High automatability means value is mostly created by processing public/standardized information and/or repeatable tasks that AI frontier models can easily replicate."
The second factor evaluates how difficult or costly it would be for competitors or customers to displace an incumbent product once it is in use. Bernstein describes this dimension in terms of product defensibility - the degree to which switching costs, ecosystem strength, embedded workflows, brand and community create a barrier to replacement. The analysts write, "Low defensibility means low switching cost, limited ecosystem, and weak lock-in/network effect. High defensibility means high switching costs, strong ecosystem, embedded enterprise workflows, and powerful brand/community."
Bernstein places emphasis on client switching cost within the defensibility dimension, noting that migration complexity, retraining, re-implementing integrations, and perceived operational risk all feed into the expense and difficulty of moving away from an incumbent supplier.
Applying this two-factor lens, the analysts identify several enterprise software names that they view as having a more favorable AI risk profile relative to many peers. The firms highlighted are SAP, Dassault Systemes and Nemetschek, which Bernstein says could be better insulated from AI disruption and would have "significant upside potential" if they successfully execute on their AI strategies.
Within the IT services cohort, Bernstein sees France's Alten and Italy's Reply as among the least exposed to competition from AI models. Capgemini is described as "relatively well-positioned," with its deep enterprise relationships and complex application and infrastructure footprints serving as protective elements.
The research note underscores a strategic imperative for services providers: a swift transition to an AI-first operating model. The analysts argue that service firms must reinvent business models rapidly to capture the shift toward services-as-software and to monetize AI effectively.
Separately, the article includes a prompt for investors evaluating value opportunities: it asks whether CAPP is currently a bargain and references a Fair Value calculator that combines multiple valuation models to assess stocks, including CAPP.
Summary
Bernstein's framework assesses AI disruption risk across software companies using two axes - automatability and product defensibility - and identifies select enterprise software and services names that appear relatively well protected. The note also stresses that services firms need to move quickly to an AI-first model to capitalize on the shift to services-as-software.
Key points
- Bernstein evaluates software AI risk using "automatability" and "defensibility" as the primary metrics.
- SAP, Dassault Systemes and Nemetschek are cited as having more favorable AI risk profiles among software vendors.
- In IT services, Alten and Reply are viewed as least exposed, while Capgemini benefits from deep enterprise ties and complex estates.
Risks and uncertainties
- Rapid improvements in AI model capabilities could raise automatability for currently protected software functions, increasing disruption risk for companies whose value relies on repeatable tasks - this primarily affects the software sector.
- Low product defensibility - including low client switching costs and weak ecosystems - would make it easier for AI-enabled alternatives to displace incumbents, affecting both software vendors and IT services firms.
- Services providers that do not transition quickly to an AI-first model may fail to capture shifts toward services-as-software, creating competitive and revenue risks for the IT services sector.