Despite fears that artificial intelligence (AI) spells the end of software, evidence suggests the opposite for enterprise incumbents. AI is reshaping how software is built and used, but it does not erode the core moats of leading platforms: systems of record, workflow ownership, proprietary data, compliance, deep industry knowledge and trust. Companies such as SAP, Sage and RELX are embedding AI to enhance automation, insight and productivity, driving faster growth rather than disintermediation. High switching costs, regulatory demands and existing customer relationships further protect incumbents. As with past innovations, AI is expanding software’s total market, acting as an additive layer, not a replacement.
Clearly, we don’t have all the answers, and the market is evolving quickly, but we believe the idea that AI represents the “death of software” is deeply wrong. The evidence, historical precedent, and current trading all suggest the opposite. Many of the so-called AI losers (SAP, Sage, RELX) are seeing accelerating revenue growth and expanding opportunities from AI, not disintermediation.
The reason incumbent enterprise software companies are well positioned is simple: while AI materially changes how software is built and used, it does not undermine the fundamental sources of enterprise software value, systems of record, workflow ownership, proprietary data, compliance, industry knowledge and trust.
The reasons we think the incumbent enterprise software companies are well placed are:
- Enterprise software is a system of record: Its core function is to define the official source of truth for an organisation, including how data is created, stored, changed, and audited. AI can assist with analysis and decision-making, but it does not replace the need for a trusted place of record that supports accountability and traceability.For example: SAP’s Enterprise Resource Planning (ERP) is not a user interface; it is the operational backbone of large enterprises. AI embedded into SAP (planning, forecasting, anomaly detection) increases automation, but the source of truth remains the ERP. Replacing SAP would require rebuilding decades of process logic, regulatory alignment, and integrations, not simply deploying an Large Language Model (LLM).
- Workflow ownership: Incumbents also own end-to-end workflows that govern how work is executed and controlled. AI can optimise individual steps, but it does not replace the orchestration, controls, and exception handling embedded in these workflows. These workflows have been built with deep industry/domain knowledge and are all interconnected. AI can accelerate individual tasks within these workflows, but orchestration, controls, and exception handling remain embedded in the core enterprise software.
- Data ownership is another key moat: Enterprise platforms control structured, permissioned, and often regulated datasets with clear provenance. AI derives value from this data but cannot substitute for ownership, governance, or access control. RELX, for example, owns proprietary datasets with domain-specific taxonomies and workflows that a generic AI model could never replicate.
- Deep industry and customer knowledge: Companies such as SAP, Sage, and RELX have accumulated decades of domain expertise, regulatory insight, and customer-specific understanding that is embedded in their products. This knowledge shapes data models, workflows, and controls in ways that are difficult for AI or new entrants to replicate.
- Regulatory compliance and risk management further favour incumbents: Enterprises must meet strict requirements around auditability, explainability, security, and data residency. AI introduces new risks including hallucinations, increasing reliance on vendors that already operate within regulatory frameworks and can provide contractual accountability.
- Switching costs remain high: Replacing core systems involves operational disruption, regulatory re-certification, retraining, and integration rebuilds. AI does not lower these costs and often raises them by increasing integration depth.
- Enterprise software is infrastructure, not an interface: Platforms such as SAP, Sage, and RELX are embedded in core operations. AI enhances these platforms but does not displace their role as foundational infrastructure.
Incumbents are structurally advantaged
They already own customer relationships, distribution, data, and trust, allowing them to embed AI quickly and at scale. Customers are more likely to consume AI through vendors they already rely on, reinforcing pricing power and platform stickiness. RELX’s legal software division has accelerated growth from 2-3% to 7-10% following the launch of its AI tools. SAP and Sage are also seeing strong uptake of AI agents and workflow tools they control, with clear monetisation opportunities given their ownership of both the agents and the data.
In summary, AI is an additive layer, not a replacement. It expands the value and relevance of enterprise software rather than rendering it obsolete.
Some simpler software tools will face greater competition from AI-native players. But for large enterprise software platforms, AI is far more likely to be a net positive than a negative. Historically, innovations such as cloud, SaaS and analytics have expanded the software market’s total addressable opportunity. AI will make coding cheaper, but the moat of enterprise software vendors extends well beyond software code alone.
