AI will not lift all European software and services alike; outcomes will depend on platforms, IP, and defensible workflows

Overview

This report summarizes Bernstein's first-half AI research on global software, European technology software, and European IT services. The core view is that AI has already begun reshaping growth, pricing, and margins in the software and services industries, but it will not benefit all companies simultaneously. The market still underestimates the speed of AI-driven differentiation: companies with platforms, data, deep workflows, and reusable IP are more likely to convert AI into higher ARPU, customer retention, and operating leverage; labor-intensive, less defensible, or highly automatable businesses may instead face revenue pool compression and pricing pressure.

Core views

The report argues that the value layer in software is shifting from traditional "systems of record" to AI-driven "systems of action," and enterprise customers are willing to pay for capabilities that can directly automate finance, supply chain, human resources, and business processes. In IT services, AI is moving delivery models from billing by person-days toward outcome-oriented and softwareized delivery, driving the rise of Services-as-Software. AI's impact is not simply divided by company labels, but depends on revenue mix, business automatability, defensibility, switching costs, data control, and execution speed. Therefore, the investment conclusion is that stock selection matters more than sector allocation.

Analysis framework

The report uses multiple proprietary frameworks to assess AI impact, including an AI leverage scorecard, an automatability-versus-defensibility map, a platform positioning framework, a Services-as-Software business model re-rating framework, and observation of FDE intensity and productization speed. The analytical focus is not how many AI features a company has, but whether AI can create sustainable pricing power, reduce delivery costs, improve retention, expand the serviceable market, and ultimately be reflected in margins.

Methodology notes

• AI business model assessmentAI leverage scorecard

  Assesses AI's real leverage on growth, margins, cash flow, and customer retention.

  This framework focuses on whether vendors can convert AI from feature demonstrations into higher ARPU, bundled packaging, shorter implementation cycles, and lower costs, rather than merely counting the number of AI product launches.

• AI impact mappingAutomatability and defensibility framework

  Explains AI risk through business automatability and product or customer relationship defensibility.

  Businesses that are highly automatable and weakly defensible are more likely to face pricing pressure and share loss; ERP, databases, and deeply embedded workflows are more resilient due to data, regulation, integration complexity, and switching costs.

• Platform positioningPlatform and workflow control-plane framework

  Judges whether a software company can upgrade from the application layer to an AI control plane.

  The report emphasizes that value will flow to companies that can embed AI into end-to-end workflows, unify data, control customer processes, and scale reuse through ecosystems.

• Service business modelsServices-as-Software re-rating framework

  Treats the migration of traditional labor services toward outcome-based, AI-delivered, and softwareized recurring revenue as a structural change.

  This framework compares traditional IT services with softwareized services across growth, gross margin, operating margin, pricing model, and customer budget migration.

• Enterprise AI deliveryFDE intensity and productization speed framework

  Uses whether Forward Deployed Engineering investment declines as customers mature to judge the scalability of AI software.

  FDE can accelerate early deployment, but if engineer input expands linearly with ARR, the model looks more like consulting services; a truly strong platform should convert project experience into reusable products, templates, and partner delivery capabilities.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• SAP

  Core beneficiary in platform ERP and supply chain software, rated Outperform with a target price of €276 / $323.

  Strengths

  ERP and supply chain are deeply embedded in customer processes, with S/4 migration, unified data, AI workflows, and high switching costs reinforcing the moat.

  Weaknesses

  AI benefits still need to be realized through execution, pricing, and product packaging.

  Comparison

  Relative to less defensible software, SAP looks more like an AI-enhanced platform than an application displaced by AI.

  Risks

  If AI investment, policy adjustments, or customer adoption speed fall short of expectations, margin and valuation realization may be constrained.

• Capgemini

  Representative beneficiary of AI-enabled services and Services-as-Software, rated Outperform with a target price of €208.

  Strengths

  Has consulting, industry expertise, AI delivery, and proprietary IP, and is well positioned to convert traditional services into a more scalable outcome-oriented model.

  Weaknesses

  In the short term it remains affected by the traditional IT services cycle, customer budgets, and transformation investment.

  Comparison

  Compared with labor-intensive service providers, Capgemini is closer to an "AI transformer."

  Risks

  If softwareized delivery advances more slowly than expected, or pricing pressure emerges before cost declines, margins may come under pressure.

• Indra

  A Best Idea driven by AI, defense, and aerospace, rated Outperform with a target price of €69.

  Strengths

  Defense demand, AI enablement, and the aerospace business enhance re-rating potential, with relatively strong demand visibility.

  Weaknesses

  The investment case is relatively sensitive to project execution, government budgets, and the pace of business re-rating.

  Comparison

  Compared with general IT service providers, Indra benefits from differentiated demand tied to defense and sovereign technology themes.

  Risks

  Government order timing, project delivery, and expectation risk after a rapid valuation increase.

• Dassault Systèmes

  Highly defensible platform software, rated Outperform with a target price of €29.

  Strengths

  Deeply embedded in engineering and design workflows, with specialized data and complex customer processes enhancing defensibility.

  Weaknesses

  The pace of AI monetization may be more gradual than at pure AI narrative companies.

  Comparison

  Within the AI impact framework, it is closer to a resilient compounding asset than highly automatable, weakly defensible software.

  Risks

  If customer IT budgets weaken or AI features fail to support incremental pricing, upside in valuation may be limited.

• Microsoft、Oracle、Workday

  Global platform software and ERP-related names, all rated Outperform.

  Strengths

  Microsoft and Oracle are seen as AI compounders that convert AI into ARPU and margin gains earlier; Workday has a foundation in HR and finance workflows.

  Weaknesses

  Names such as Workday still rely more on validation of AI execution and customer adoption.

  Comparison

  Compared with smaller generalist software vendors, scaled platforms have advantages in data, R&D, cloud ecosystems, and cross-selling.

  Risks

  Commoditization of AI features, rising compute costs, insufficient customer willingness to pay, or regulatory pressure.

• Sage、Nemetschek、Reply、TeamViewer、Worldline

  Market-Perform group, reflecting coexistence of AI opportunities and business constraints.

  Strengths

  Sage has defensibility in mid-market ERP, Nemetschek has some moat in professional workflows, and Reply has premium consulting and engineering characteristics.

  Weaknesses

  Some companies are weaker than large platforms in AI monetization, scale, and end-to-end platform control.

  Comparison

  Better suited for selective holding rather than indiscriminate chasing of the AI theme.

  Risks

  Insufficient pricing power, AI investment costs, volatility in vertical demand, and intensifying competition.

• Atos、CGI and labor-intensive IT service models

  The report remains cautious on labor-led, weakly defensible, or highly automatable businesses; Atos and CGI are rated Underperform.

  Strengths

  Existing contracts and customer relationships may provide some short-term stickiness.

  Weaknesses

  Time-and-materials billing, offshore labor arbitrage, and standardized processes are more vulnerable to compression from AI.

  Comparison

  Compared with Services-as-Software leaders, traditional service models are less able to achieve software-like margins and valuation multiples.

  Risks

  Contract renegotiation, pricing pressure, client insourcing, automation substitution, and shrinking revenue pools.

Key data

• Enterprise willingness to replace servicesAbout 60% of enterprises plan by 2028 to replace labor-led services with softwareized deliveryUsed to support that Services-as-Software is not a cyclical recovery, but a structural transformation in delivery models.
• AI-enabled ERP penetrationClose to about 90% by 2030AI is pushing ERP from a system of record toward a system of action.
• Mid-market ERP spending growthAbout 13% CAGR by 2030Cloud migration still provides stability, but the next phase of growth will depend more on AI deployment, pricing discipline, and value realization.
• Cumulative supply chain software opportunityAbout $217bn cumulatively from 2026 to 2029The report believes the market underestimates the scale and persistence of the AI-driven supercycle in supply chain software.
• Supply chain software market sizeAbout $33bn in 2024, about $66bn in 2029, about 15% CAGRGrowth is driven by resilient supply chains, geopolitical restructuring, ESG compliance, customer service requirements, and AI adoption.
• AI-enabled SCM growthAbout 46% annual growth through 2030, with AI-enabled spend approaching about 80% by 2030AI is moving from decision support toward autonomous execution, bringing higher investment intensity and ROI.
• SCM ROIAbout 25%-30%Driven by optimization in inventory, logistics, procurement, forecasting, and service levels.
• SaaS SCM growth and cloud penetrationSaaS SCM about 23% annual growth, with cloud penetration rising from 52% in 2024 to 74% in 2029Cloud migration and outcome-based pricing form the second monetization path.
• Services-as-Software market opportunityCurrently below $100bn, about $600bn-$700bn by 2028, and up to about $1.5tn in the long termThis market will cannibalize both traditional IT services and SaaS budgets.
• Services-as-Software margin assumptionsGross margin 65%-70%, operating margin 15%-25%Higher than traditional services at about 25%-30% gross margin and 5%-15% operating margin.
• Supply chain software market concentrationTop three hold about 38% share, with SAP at about 25% shareEnd-to-end platforms, ERP integration, and M&A may strengthen the advantages of leaders.

Impact & implications

The investment implication is a shift from sector beta to company alpha. Companies with strong AI monetization, strong control over data and workflows, and the ability to productize service projects are more likely to achieve valuation re-rating; companies whose revenue depends on labor, whose delivery is not scalable, or whose businesses are easily displaced by automation will face pricing pressure, margin dilution, and market share risk. The report therefore emphasizes owning platform software leaders and IP-driven service leaders, while avoiding the simplistic view that AI is a blanket positive for the entire industry.

Risks

• AI features may become table stakes before pricing benefits are realized, leaving software vendors to bear higher R&D and compute costs.
• Smaller or non-scaled software vendors may lack advantages in cloud costs, data, R&D, and ecosystems, facing margin dilution.
• Highly automatable businesses such as analytics, application development, and business process services may face structural pricing pressure.
• Outcome-based pricing will increase revenue volatility and require vendors to invest in measurement, customer success, and AI infrastructure.
• If FDE expands linearly with revenue, it will weaken software scalability and make the business model look more like bespoke consulting.
• Enterprise AI implementation may be hindered by data quality, governance, lack of accountable owners, and poorly defined use cases.
• Customer dependence on vendor engineers may create short-term stickiness, but in the medium term it may also trigger budget ceilings, partner replacement, or replatforming risk.

What to watch

• Whether AI products drive ARPU uplift, better retention, pricing upside, and higher free cash flow conversion.
• Whether FDE investment per customer declines and whether project experience is converted into reusable products, templates, and partner delivery.
• The pace of adoption of agentic and autonomous workflows in ERP, and progress toward high AI-enabled ERP penetration by 2030.
• Progress in AI-enabled spend, SaaS migration, outcome-based pricing, and platform consolidation in supply chain software.
• Whether enterprise customers replace labor-led services by 2028 as planned and shift budgets toward Services-as-Software.
• Contract renegotiation, pricing pressure, workforce restructuring, and insourcing trends among labor-intensive service providers.
• Whether AI monetization evidence at core names such as SAP, Capgemini, and Indra is sufficient to support further valuation differentiation.