Artificial Intelligence (AI) is rapidly emerging as a transformative force, disrupting various industries and fundamentally altering the way we work. Since Generative AI achieved a major enterprise breakthrough in 2025, organisations have been steadily moving from experimentation to large-scale adoption across business functions.
Why Enterprise Architecture Management matters for Agentic AI
Agentic AI represents the next evolutionary step in the AI journey – moving from systems that merely assist humans to autonomous, goal-driven agents capable of reasoning, acting, and collaborating across enterprise environments. While organisations around the globe are embracing Agentic AI in their AI transformation journey, only 11% of German companies have achieved higher revenues and 16% have reduced costs from AI over the past 12 months, leaving them behind their global peers (Source: 29th PwC Global CEO Survey). Enterprise Architecture Management plays a critical role in changing that as it provides the structure, tools, and governance not only to explore, but also to scale Agentic AI at the right pace and scale, so organisations can generate value quickly and avoid being overtaken by competitors.
As businesses navigate an evolving landscape, Agentic AI brings both challenges and opportunities for Enterprise Architecture (EA) and its Management discipline (EAM).
First, leveraging AI agents at scale often requires re-architecting an organisation’s technology foundations to fully realise the potential of Agentic AI. Second, Agentic AI changes ways of working and demands new skill sets for developing, managing, and governing AI-enabled systems. As AI becomes more autonomous and deeply embedded in business processes, EAM is essential to align AI capabilities with business strategy, govern complexity, and balance business opportunities with risks related to security, compliance, and ethical use. Third, Agentic AI augments human analysis and decision-making, paving the way for greater maturity in the organisation’s EAM capability by increasing its efficiency, influence, and business value.
To harness these benefits, leaders are under pressure to prepare their business and IT architecture as well as their EAM practice for Agentic AI adoption.
Re-architecting technology foundations for Agentic AI
Creating foundations for Agentic AI requires re-architecting enterprise technology platforms and architectures capable of supporting autonomous agents at scale (Gartner Top 10 Strategic Technology Trends for 2026):
At the core, this includes high-performance, hybrid computing infrastructure that can handle the intensive processing demands of agentic AI workloads – from real-time reasoning to continuous learning. Meeting these demands also requires a robust data architecture that provides standardised, secure, low-latency access to structured and unstructured data, while supporting real-time ingestion, transformation, and retrieval.
Equally important are integration and orchestration frameworks that enable AI agents to interact seamlessly with existing business applications, external services, and other AI-driven processes via APIs, event-driven pipelines, and service meshes, with support for open standards such as the Model Context Protocol (MCP). Embedded governance and security mechanisms – including hardware-based confidential computing and trusted execution environments (TEEs) – ensure that sensitive data, AI models, and operational processes remain protected, auditable, and compliant with regulatory requirements.
Finally, AI-native development environments and tooling accelerate experimentation, deployment, and scaling of agentic workflows, allowing teams to iterate safely and efficiently while reducing reliance on large, centralised development teams. Together, these layers form a resilient, flexible, and secure foundation that enables organisations to operationalise Agentic AI safely, efficiently, and at scale.
Because these capabilities are prerequisites for the enterprise adoption of Agentic AI, EAM must evolve from a supporting function into a strategic enabler that effectively guides the change initiatives required to re-architect the technology foundations.
Developing, managing, and governing Agentic AI
Meanwhile, organisations must fundamentally rethink how they develop, manage, and govern an IT landscape increasingly augmented by Agentic AI.
As AI agents transition from isolated experimental components to integrated, productive building blocks of the enterprise, management practices across the entire IT lifecycle must be adapted.
From a development perspective, AI agents become integral parts of digital products, whether developed in-house or embedded in off-the-shelf applications and services.
Organisations must establish robust processes for how AI agents are developed, integrated, tested, and operated within an existing IT landscape, while ensuring that knowledge, principles, and best practices are shared across teams rather than being concentrated within a small group of specialists.
Within IT management, AI agents represent a new IT asset class, fusing the modularity and autonomy of microservices with “intelligent” capabilities such as cognitive reasoning, proactivity, and interaction via natural language prompts. This new asset class demands that IT management practices be revised. For example, clear ownership of AI agents and accountability for their actions must be defined, data and knowledge considerations must be integrated into lifecycle management, and audit and approval processes must address AI-specific risks.
Because AI agents operate with greater autonomy and may directly influence business decisions, traditional governance models are insufficient. Effective governance for Agentic AI must incorporate specific dimensions, including data governance, principles for ethical and responsible AI, transparency, and explainability. This also requires robust controls to ensure the quality and accountability of AI-driven outcomes, such as key management decisions.
By coordinating the evolution of development, management, and governance, EAM enables the organisation to safely embed and confidently scale Agentic AI while maintaining control, trust, and alignment with business objectives.
Leveraging Agentic AI for EAM
EAM plays a fundamental role in ensuring that the introduction and scaling of AI agents are controlled, transparent, and value-driven. Building on the technology foundations and new ways of developing, managing, and governing the AI-augmented IT landscape, organisations can reimagine how their EAM capability can be enhanced with Agentic AI.
EAM processes and services that were previously manual, expert-driven, or simply too complex to automate are prime candidates to explore the automation potential of Agentic AI.
For example, business users can use an AI agent’s chat interface to request new applications or services, reducing recurring portfolio discussions about existing alternative solutions. Further, EAM’s role in IT demand management can be augmented by Agentic AI to evaluate architectural changes arising from demands and projects, including impact analysis and compliance checks based on knowledge in the organisation’s architecture repository. At the same time, EAM teams can be augmented by specialised AI agents with access to curated architectural knowledge sources and tools, enabling more efficient modelling, deeper insights, and more consistent decision-making.
To realise this potential, organisations should proactively explore and leverage AI capabilities already embedded in their existing tool landscape – such as AI-enabled EA platforms and services – to integrate them into day-to-day EAM workflows.
Kick-start your Agentic AI journey
Agentic AI is rapidly becoming a priority for decision-makers, but for most organisations, tangible near-term benefits remain unclear. EAM provides the structure, tools, and governance to explore and scale Agentic AI across the enterprise and turn adoption into value and competitive advantage. Accordingly, both the EA – the business and IT architecture – and the EAM function must evolve to build the required capabilities to the right extent and in the right sequence: establish the technology foundations to integrate AI agents, rethink development, management, and governance for an AI-augmented IT landscape, and leverage Agentic AI to enhance EAM processes and services.
Our team at PwC Germany brings deep expertise in digital transformation, EAM, and AI agents. We help you identify and implement high-value Agentic AI use cases today and craft strategic roadmaps for future transformation – aligning speed with safety and compliance.
To learn how to build Agentic AI foundations, assess your organisation’s maturity, and decide where and how to apply Agentic AI, get in touch with our team.
Die Autoren
