Agentic AI is reshaping enterprise software moats: value is shifting from application interfaces to orchestration and context layers

Overview

This report discusses the impact of Agentic AI on moats, business models, and valuations in the European enterprise software sector. The core judgment is that AI will not simply replace application software across the board, but will instead reallocate value across the technology stack: core applications, systems of record, regulation, and workflow context retain defensive value, while higher incremental value will be contested in the agent, orchestration, context data, and intelligence layers. Current European software valuations already reflect part of the structural disruption risk, but whether the sector can bottom depends on evidence of AI product delivery, adoption, monetization, pricing power, and margin resilience.

Core views

The report's core views include: first, enterprise software interaction will shift from humans using software to agents calling SaaS, raising the importance of APIs and central orchestration layers. Second, systems of record will not lose value because of AI; instead, they become even more critical due to training, permissions, audit, governance, and contextual execution needs. Third, the true moat is not raw data itself, but workflows, metadata, permissions, business logic, audit capabilities, and cross-silo coordination built over many years. Fourth, vertical software and mission-critical horizontal software are relatively more defensive, while thin wrappers, single silos, pure dashboards, and low-switching-cost software are more vulnerable. Fifth, AI business models will drive an evolution from seat-based pricing toward hybrid pricing based on consumption and outcomes, while creating near-term gross margin pressure.

Analysis framework

The report analyzes the issue through four lenses: technology stack restructuring, software moat decomposition, business model economics, and valuation cycle comparisons. On the technology stack side, it distinguishes front-end interfaces, agents, orchestration, context layers, systems of record, data platforms, and cloud infrastructure. On the business model side, it compares traditional SaaS seat-based pricing with the variable costs of the AI era driven by tokens, inference, and workflow complexity. On the valuation side, it combines share-price drawdowns among covered European software companies, reverse-DCF implied growth, historical technology cycles, and company ratings to assess conditions for re-rating.

Methodology notes

• Technology stack analysisAgentic AI enterprise software stack

  Enterprise AI will be composed jointly of multiple specialized agents, central orchestration, a context data layer, systems of record, and underlying infrastructure.

  This framework is used to judge where value will be captured: foundational systems provide data, governance, and workflows, while upper-layer agents and orchestration compete for incremental intelligence value.

• Moat analysisContextual layer moat

  Context-layer moat

  The context layer includes processes, workflows, metadata, permissions, audit, embedded business logic, and cross-system coordination; the report argues it is harder to replicate than extractable raw data.

• Business model analysisHybrid monetisation model

  Seat-based, consumption-based, and outcome-based pricing coexist

  AI inference and token costs make software revenue and costs more variable, potentially weakening the predictability and high-gross-margin advantages of traditional SaaS.

• Valuation analysisReverse DCF and historical cycle comparison

  Reverse DCF and historical technology cycle comparison

  The report uses growth implied by current share prices, margin assumptions, WACC, and comparisons with historical cycles such as the cloud transition to assess whether software valuations are near historical troughs.

Asset mapping & comparison

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

• SAP

  One of the most defensive names in coverage, rated Buy.

  Strengths

  Mission-critical enterprise workflows, cross-domain integration, and deep accumulation in systems of record and context layers; Joule can serve as an orchestration entry point for agents to access data and context.

  Weaknesses

  Whether its AI orchestration role can be fully monetized remains to be proven.

  Comparison

  Compared with more horizontal or shallow applications, SAP is more deeply embedded in systems of record and enterprise processes.

  Risks

  AI product rollout, customer adoption, and monetization progress may fall short of expectations; third-party agent layers may capture incremental value.

• Nemetschek

  Relatively well positioned, rated Buy.

  Strengths

  Vertical software in architecture, deep industry expertise, and embedded workflows provide defensiveness.

  Weaknesses

  Vertical customers may build some AI capabilities themselves through LLMs, system integrators, or next-generation data providers.

  Comparison

  It has greater domain complexity and switching costs than horizontal tools.

  Risks

  A stronger build-versus-buy trend may constrain upsell and cross-sell opportunities.

• Dassault Systemes

  Relatively well positioned but rated Neutral.

  Strengths

  Deep manufacturing domain expertise and complex workflows provide a moat.

  Weaknesses

  AI-native or customer-built capabilities may compete for upper-layer intelligence value.

  Comparison

  More defensive than thin-front-end and simple horizontal tools, but valuation and monetization evidence still need to be observed.

  Risks

  Uncertainty around value migration to the agent layer, product roadmap execution, and client budget reallocation.

• Temenos

  Benefits from its highly regulated industry exposure, rated Neutral.

  Strengths

  Compliance, audit, permissions, and systems-of-record requirements in financial software increase customer stickiness.

  Weaknesses

  Growth and re-rating still depend on its AI roadmap and monetization capability.

  Comparison

  Its regulatory characteristics make it more defensive than simple CRM-like or generic horizontal software.

  Risks

  Customers may use third-party AI or system integration solutions to build upper-layer functionality, weakening pricing power.

• Sage

  Relatively more exposed, rated Neutral.

  Strengths

  Accounting software has some regulatory characteristics, providing partial defensiveness.

  Weaknesses

  Its horizontal software positioning and SMB customer mix make it more vulnerable to AI substitution and embedded alternatives.

  Comparison

  Compared with SAP or deeply vertical software, it has weaker business context and enterprise embedding.

  Risks

  AI agents executing a broad range of tasks more efficiently may compress seat demand and upsell opportunities.

• TeamViewer

  Rated Neutral.

  Strengths

  It has an existing customer base in certain enterprise use cases.

  Weaknesses

  The report table shows current reverse-DCF implied growth at -5.5%, reflecting low market expectations for long-term growth.

  Comparison

  Under the report's framework, it is not among the clearest beneficiaries of AI moats.

  Risks

  Value migration toward AI orchestration, automation, and platform layers may limit growth.

• Sinch AB

  Rated Sell.

  Strengths

  It has an existing foundation in communications software.

  Weaknesses

  The FY26 margin assumption is relatively low, disclosed in the table at 13%, while reverse-DCF implied growth is 12.5%.

  Comparison

  Compared with enterprise software companies that have deep systems of record and complex workflows, its defensiveness is weaker.

  Risks

  Competition, AI substitution, margin pressure, and demanding valuation-implied growth expectations.

Key data

• European software coverage share-price drawdownabout 37%Covered companies' share prices are down about 37% on average from their 52-week highs.
• AI-era gross margin assumptionabout 60%-80%The report believes AI hardware and inference costs will pressure gross margins in the near term.
• SaaS-era gross margin rangeabout 75%-85%Used as a comparison benchmark for AI-era gross margin pressure.
• On-premise era gross margin rangeabout 85%-95%Used by the report to compare structural gross-margin changes across software model transitions.
• Token cost changedown more than 90% over the past few yearsBut usage has expanded faster, making cost governance and model routing still important issues.
• SAP reverse-DCF implied growth4.3%The table shows SAP's current share price at 144.0 in local currency, FY26 margin at 30%, WACC at 8.3%, and rating at Buy.
• Nemetschek reverse-DCF implied growth5.1%The table shows a current share price of 57.9, FY26 margin of 32%, and WACC of 10.0%.
• Sinch reverse-DCF implied growth12.5%The table shows a current share price of 42.41, FY26 margin of 13%, WACC of 13.4%, and rating at Sell.

Impact & implications

For investment judgment, the AI shock should not be interpreted simply as a wholesale failure of software incumbents, but rather through distinctions in value-chain position and moat quality. Companies with mission-critical processes, strong regulatory requirements, cross-domain data context, and monetizable orchestration capabilities are more likely to retain or re-expand value; companies that rely on thin front ends, low-differentiation interfaces, single data silos, or functionality that customers can easily build themselves are more likely to face pricing-power and growth pressure. Whether the sector can re-rate will require evidence of actual AI product delivery, deployment of agent-based workflows, monetization paths, customer acceptance of consumption-based pricing, and gross-margin durability.

Risks

• AI product adoption may be high, but customer willingness to pay may be insufficient, leaving monetization paths unclear.
• Token, inference, and hardware costs may rise faster than model prices fall, compressing software company gross margins.
• Third-party agents, LLM vendors, and system integrators may capture upper-layer intelligence and orchestration value.
• Customers may build AI applications themselves on top of core systems, weakening incumbents' upsell opportunities and pricing power.
• Thin wrappers, single silos, and pure front-end analytics and dashboard software may be bypassed by APIs and agent layers.
• Current valuations may be repricing terminal-value risk rather than merely reflecting short-term earnings pressure, making it difficult to confirm an industry bottom.

What to watch

• Whether incumbents' Agentic AI product roadmaps are delivered on schedule.
• Enterprise rollout, actual usage frequency, and retention of agent-based AI solutions.
• Whether customers are willing to pay incremental fees for orchestration, context layers, governance, and outcome-based functionality.
• The impact of consumption-based and outcome-based pricing on revenue growth, predictability, and customer budgets.
• Whether AI inference, token, and hardware costs continue to depress gross margins, or can be offset by opex optimization.
• Whether the scope of customer-built AI capabilities in vertical software expands from peripheral features into core workflows.
• Evidence of valuation re-rating for SAP, Nemetschek, Dassault Systemes, Sage, Temenos, TeamViewer, and Sinch.