On 1 June 2026, the Architecting the Autonomous Enterprise webinar series officially launched with its opening session together with Mark Smith and Paul Kurchina,  where they provided a strategic overview designed to set the foundation for sessions that follow!

This series is built by architects, for architects. Running weekly, from 1st June through to 20 August, it is designed to help Enterprise Architects make sense of the announcements coming out of SAP Sapphire 2026 from the SAP Business AI Platform and Joule Work to the SAP Autonomous Suite, Knowledge Graph, AI agents, and governance — and understand what they mean in practice for enterprise architecture, data foundations, and transformation.

Register here to get access to the slides recordings and join the following sessions: https://events.sap.com/eaa-autent-web/en_us/home.html

“The Business AI Platform and Why Foundations Come First”

– Mark Smith

Mark Smith set the strategic context for the series, drawing directly from conversations with customers and partners at SAP Sapphire in Orlando and Madrid.

The Problem: Enterprise AI Is Stuck

Mark opened with a clear framing of the current reality. Three-quarters of enterprises running AI pilots have no clear path to scale. Nearly half cannot get past data quality and access issues. The investment is real, but the results are not. The root cause, he argued, is what he calls the context deficiency — three gaps that keep enterprise AI stuck in pilot mode:

1. Business process context. Generic models can read data, but they cannot reason over how a business actually runs. They see tables, not operations.
2. Data connection. AI is layered on top of disconnected, outdated systems and asked to be intelligent about a business it can only see in fragments.
3. Governance. AI that cannot be audited or controlled does not get deployed into the core of the enterprise. It gets demonstrated. There is a difference.

SAP's Answer: Three Foundations Already in Place

Mark then laid out SAP's structural differentiation, not as a roadmap, but as deployed capability:

1. Deep Process and Industry Knowledge. Fifty years of business process expertise encoded in SAP applications, the Knowledge Graph, and purpose-built models. This is not a marketing number — it is the reason SAP's models deliver 3.5 times higher prediction quality than generic LLMs, grounded in 7.3 million data fields covered by the Knowledge Graph.
2. Semantically Rich Business Data. SAP data carries business meaning by default. A customer is linked to a contract. A material is tied to an order. A supplier is scored for risk. AI reasons over real business context across both SAP and non-SAP systems.
3. Enterprise-Grade Governance. SAP has been in governance, risk, and compliance longer than most AI companies have existed. Governance is not a feature added on top it is foundational to the architecture.

The SAP Business AI Platform

Mark introduced the SAP Business AI Platform launched at Sapphire as the foundation for how enterprises will build, connect, and run AI that truly understands their businesses. It is organised around three pillars:

• Build. A single unified environment for building AI agents, applications, and workflows grounded in business context — using low code or pro code, fully integrated with SAP and non-SAP systems. Joule Studio, the SAP Integration Suite, and integrations with tools like Claude and Cursor enable developers to build from where they are today.
• Contextualize and Reason. Only SAP can ground AI in fifty years of business process expertise plus the millions of hours organisations have invested in their own implementations. The data fabric serves as the trusted knowledge core, connecting data, processes, and semantics so agents reason accurately and avoid hallucinations.
• Govern. Scale without control creates exposure. Governance is not a layer on top but the command centre for enterprise-wide data and AI agent governance — from planning through production, with full auditability, lifecycle control, and secure execution.

The Autonomous Enterprise Vision

Mark closed his segment by framing the Autonomous Enterprise as SAP's North Star: the future where AI transforms how people work and how processes run by grounding decisions in real-time intelligence, automating end-to-end workflows, and proactively improving every function. Five elements work together to make it real:

1. Joule : the engagement layer and front door, where users express intent and Joule orchestrates agents and systems to make it happen.
2. The SAP Autonomous Suite : the operational core, running the business across five domains: Finance, Spend, Supply Chain, HCM, and Customer Experience.
3. Industry AI : deep industry intelligence bringing sector-specific process logic, data context, and regulatory rules into AI-powered solutions tailored to each industry.
4. The SAP Business AI Platform : the foundation providing context, unified data, reasoning, and governance in one place.
5. Agent-led transformation : the path from where organisations are today to the autonomous model, without stopping the business to transform the business.

The message to architects: people set the direction, AI executes. This is the shift from recording what happened to helping execute what is next.

 Register here to get access to the slides recordings and join the following sessions: https://events.sap.com/eaa-autent-web/en_us/home.html

The Context System and the Architect's Starting Point

I took second part of the session deeper into architectural territory, introducing a framework that will underpin much of the series: the Context System.

Why Context Is Everything

We framed the fundamental shift: this is not task automation. Organisations are now delegating decision-making and judgment to systems that operate through language-like capacity. The critical differentiator is whether your solution has the right context, not just data, not just process information, but a complete context system that includes all the deviations, exceptions, and organisational knowledge that define how work actually gets done. SAP's fifty years of running ERP systems for the world's largest organisations has produced exactly this: a right to win built on deep, operational context available across the full stack.

Five Layers of the Context System

We unpacked the Context System as five distinct architectural layers as written by John Santic in his detailed writeup on The Context System.

1. Data Context — What Happened The foundational layer: transactions, records, orders, invoices, inventory states, financial postings. All post-facto — the transaction has already occurred. SAP Business Data Cloud, combined with the Demeo acquisition (Apache Iceberg capabilities) and BDC Connect, forms the starting point for this layer.

2. Semantic and Knowledge Context — What Does It Mean Data tells you what happened. This layer tells you what it means. Customer versus account versus ship-to — an agent does not inherently understand the difference. The Knowledge Graph and SAP's acquisition of WalkMe (referred to as “Rails” in the session) provide entity understanding, master data relationships, and lineage across systems.

3. Process and Predictive Context — How Work Gets Done The exact process followed by people and systems: approvals, handoffs, bottlenecks, deviations from the intended path. SAP Signavio provides the process intelligence, while Prior Labs' tabular foundational models add predictive capabilities. This is also where the concept of Agent Mining emerges understanding the exact behaviour of agents within systems to enable self-learning and improvement.

4. Organisational and Governance Context — What Is Allowed Today, there is no standard way of governing across different agents. The SAP AI Agent Hub, built on top of LeanIX, provides the capability to govern, provision, observe, and manage every agent across the landscape — whether SAP or non-SAP. This brings the operational and governance context needed for full lifecycle management.

5. Intent and Prescriptive Context — The Company Memory The operational meaning of everything an organisation does. How humans act, and how agents need to serve alongside them. This is what Christian Klein described as the “company memory” in his Sapphire keynote — the layer that enables humans and agents to work together with full contextual understanding.

 We also highlighted four assets already available on the SAP Architecture Center for architects ready to start:

• AI-Native North Star Architecture — how SAP envisions the convergence of AI, data, and applications across the Business AI Platform, Business Data Cloud, agents, autonomous domains, and Joule as the unified engagement layer.
• The AI Golden Path — a guided path from idea to production-ready applications, including detailed decision trees for when to use agents vs. deterministic applications vs. Knowledge Graph vs. HANA Vector Engine vs. agentic workflows vs. code-based agents.
• Agentic AI Architecture Pattern — interoperability between Joule and third-party agents, including details on Agent Hub, Joule Interface, and Agent Gateway.
• Data Architecture with Business Data Cloud — how to extract data from SAP systems without moving it (zero-copy, delta share protocol), modernise BW landscapes, and build the data layer.

Paul Kurchina served as the session's connector, opening the series, framing the philosophy, moderating the audience Q&A, and linking the session to the broader community programme.

The Philosophy: Learning Out Loud Together

Paul emphasised that this series is not a broadcast, it is a shared learning journey. The Autonomous Enterprise was launched just weeks before this session, and everyone including SAP's own teams is learning and growing together. The series is designed to evolve based on participant feedback, with new sessions added as needs surface.

Live Q&A Highlights

Paul moderated four questions from the audience, answered by Mark Smith:

Q: Where is the biggest gap between where customers are today and autonomous operations? Most organisations are experimenting with AI, but when asked who is operating autonomously, 95% of hands go down. The gap is in InfoSec sign-off, understanding of MCP servers interacting with core processes, and the absence of event-driven architecture. Without AI foundations in place, organisations remain stuck in pilots.

Q: Is Clean Core a prerequisite for agent-led transformation? Clean Core makes AI scalable, but the approach must be balanced combining AI adoption with a path to simplification. AI can actually accelerate modernisation and reduce technical debt. The key is identifying which process domains can be automated and building a holistic roadmap from current state to the autonomous enterprise.

Q: What can SAP do that hyperscalers cannot? SAP's true differentiator is grounding AI in how businesses actually run — fifty years of process expertise, the Knowledge Graph, and purpose-built models like RPT 1.5 that are pre-trained with business context. Combined with fully managed governance, lifecycle control, and secure execution via the SAP AI Agent Hub, this provides what external AI providers structurally cannot replicate.

Q: What should every architect do in the next 90 days? Get the foundations in place for agentic execution. Move from pilots to repeatable patterns for enterprise execution. The Business AI Platform provides those foundational elements — and the design-time environment is free of charge, removing barriers to building, testing, and prototyping.

Register here to get access to the slides recordings and join the following sessions: https://events.sap.com/eaa-autent-web/en_us/home.html

What Comes Next

The in-person tour. In addition to the webinar series, a six-city in-person learning tour is running across North America (Palo Alto, Atlanta, Dallas, Toronto, Newtown Square, and Calgary) bringing architects together face-to-face to discuss what they are learning through the series.

The series will adapt based on participant input. Surveys follow each session, and new topics or deeper dives will be added based on what architects tell the team they need most. You can still provide feedback here if you have not done already: Architecting the Autonomous Enterprise Webinar Series – Fill out form

What's Up for the Next Session:

The next session on 10 June will focus on what SAP's evolved portfolio means for architects together with Mohamed Abdel Hadi, Habeebuddin Mohammed and Jason Luo examining how the full portfolio comes together (Joule, Joule Work, the SAP Autonomous Suite, Industry AI, and the Business AI Platform) and what it means for architecture in an agentic world. We will also cover How SAP Data and AI services help you craft the right architecture!

Thank you to everyone who joined and hope you continue to support us through this webinar series!