SAP’s Cloud and AI Momentum Is Growing — But What Should Customers Do Next?.

SAP’s Q2 2026 results tell a clear story. Cloud adoption continues to grow, and AI is becoming a much bigger part of SAP’s overall direction. SAP reported strong growth in its cloud backlog and Cloud ERP Suite revenue. That is important, but the numbers alone are not the most interesting part of the announcement. The bigger shift is this: SAP is no longer talking about cloud migration and AI as two separate topics. Cloud ERP is becoming the foundation on which SAP expects customers to build and run AI-enabled business processes. For customers, this changes the conversation. The question is no longer just, “When are we moving from ECC to S/4HANA?” It is now also, “Will our processes, data and architecture allow us to use AI effectively once we get there?” ##Cloud ERP Is Becoming the Operating System for Enterprise AI For the last several years, most S/4HANA programs have focused on familiar goals: -Moving away from aging ECC platforms -Standardizing business processes -Reducing technical debt -Simplifying custom developments -Improving reporting -Building a cleaner and more upgrade-friendly ERP landscape All of these goals are still relevant. What has changed is that SAP is increasingly positioning S/4HANA and the broader cloud portfolio as the execution layer for enterprise AI. Joule is no longer being positioned only as a conversational assistant. SAP is moving toward assistants and agents that can work across finance, procurement, supply chain, manufacturing, HR, sales and service processes. That means the future SAP experience may look very different from what users are used to today. Instead of opening several transactions, reports and inboxes, a user may simply describe the outcome they need. An agent could then: -Gather information -Compare options -Identify exceptions -Recommend a decision -Execute an approved action -Escalate anything that needs human judgment That is a significant change from the first wave of generative AI. ## AI Is Moving from Answers to Actions Most early enterprise AI use cases were focused on summarization, document generation and question answering. Those use cases still have value, but the next phase is about getting work done. Think about a few practical examples. A procurement agent could identify delayed purchase orders, check alternative suppliers and propose the next best action. A warehouse agent could review outbound priorities, available labor and inventory exceptions before recommending a work plan. A finance agent could identify billing or reconciliation issues and prepare corrective actions. A maintenance agent could review equipment history, symptoms and spare-parts availability before recommending what a technician should do next. A sales agent could check customer requirements, inventory, credit status and delivery feasibility before confirming an order. These are not simple chatbot scenarios. They require access to trusted data, clear business rules, secure APIs, proper authorizations and a well-defined approval process. That is where the real work begins. ## Moving to the Cloud Does Not Automatically Make a Customer AI-Ready This is where organizations need to be careful. A successful cloud migration does not automatically mean the organization is ready for AI agents. An agent will only be as good as the environment around it. If master data is inconsistent, the agent will make decisions using inconsistent information. If the same business process is followed differently across plants or regions, the agent may not know which path is correct. If APIs are missing or unreliable, the agent may be able to provide an answer but not complete the task. If authorizations are poorly designed, the risk becomes much more serious. An agent should never be able to perform an action that the user behind it is not authorized to perform. If there is no monitoring or audit trail, the organization may not know why an action was taken or how to reverse it. In my view, AI readiness for SAP customers comes down to five areas. ## 1. Process Readiness The process should be understood before it is automated. Customers need to know: -What the standard process is -Where exceptions occur -Which decisions follow clear rules -Which decisions still require human judgment -Where delays and rework happen today Automating a poorly understood process usually creates faster confusion, not better outcomes. ## 2. Data Readiness AI requires more than access to tables. It needs trusted data, clear business meaning and well-governed master data. A material number, customer record or equipment status may mean different things depending on the process context. The agent needs that context. ## 3. Integration Readiness Agents need a safe way to interact with SAP. That may include: -Released APIs -OData services -CAP or RAP services -Events -Workflows -Governed function modules -Custom business services Without these building blocks, AI remains mostly advisory. ## 4. Security Readiness Every agent action must have a clear identity behind it. Organizations need to know: -Which user is requesting the action -Which SAP identity is being used -What that user is authorized to do -Whether an approval is required -How the action will be logged Security cannot be added after the agent is built. It has to be part of the design from the beginning. 5. Operational Readiness Agents also need to be monitored like any other enterprise system. Customers should be able to measure: -Accuracy -Completion rate -Exception rate -Human intervention -Business impact -Authorization failures -Incorrect recommendations -Reversed or cancelled actions Without these measures, it is difficult to know whether the agent is creating value. ## Clean Core Is Becoming an AI Topic Clean core is usually discussed in the context of upgrades and extensibility. That is still important, but AI adds another dimension. A clean core gives agents a clear and governed way to interact with SAP. Customers need stable APIs, well-designed services, clear business objects, proper authorization checks and auditable workflows. An environment with thousands of undocumented custom objects and point-to-point integrations will be much harder to expose safely to AI. This is why clean core should no longer be viewed only as a technical modernization initiative. It is also an AI-enablement initiative. ## SAP BTP Will Play a Major Role SAP BTP is likely to become the main layer connecting SAP applications, external systems, enterprise data and AI agents. Customers can use BTP to: -Build clean-core extensions -Create CAP and RAP services -Integrate SAP and non-SAP systems -Develop custom agents and Joule skills -Orchestrate workflows -Use SAP AI Core and the generative AI hub -Manage application and vector data -Implement identity mapping -Monitor AI interactions -Add approval and governance controls The challenge is that many customers still treat BTP as a collection of services purchased for individual projects. That approach will become difficult to manage. BTP needs an enterprise architecture, governance model and roadmap. ## Start with the Business Problem, Not the AI Tool One of the biggest mistakes organizations can make is starting with the model or platform. The better starting point is the process. Ask questions such as: -Where do employees spend time collecting information? -Which processes depend heavily on emails and spreadsheets? -Where do teams repeatedly move between SAP and non-SAP applications? -Which decisions follow consistent business rules? -Where do delays affect revenue, inventory, service levels or customer experience? -Which processes generate a high volume of exceptions? -Where does valuable knowledge sit with only a few experienced employees? Once the opportunities are identified, they can be grouped into four levels. Informational The agent retrieves and summarizes information. Advisory The agent analyzes information and recommends an action. Transactional The agent performs an approved action. Autonomous The agent completes a multi-step business process within clearly defined limits. Most customers should begin with informational and advisory use cases. That provides value while keeping the risk manageable. ### What SAP Customers Should Do Now Customers do not need to launch dozens of AI projects immediately. They do need to start preparing. A practical first step would be: 1. Identify three to five high-value business processes. 2. Measure the current effort, delays and error rates. 3. Review the required SAP and non-SAP data. 4. Identify available APIs and integration gaps. 5. Define the user and authorization model. 6. Decide where human review is required. 7. Build a focused proof of value. 8. Measure business outcomes. 9. Reuse the architecture for the next use case. 10. Put governance in place before scaling. The best opportunities are often in processes where SAP already contains the core business information, but users still spend significant time connecting the dots manually. ## How Mygo Can Help At Mygo, we work across SAP architecture, S/4HANA, BTP, integration, data, automation and AI. That allows us to approach SAP Business AI from a business-process perspective rather than treating it as a standalone technology project. ## AI and Agent Readiness Assessment We help customers assess whether their SAP landscape is ready for AI. This includes: -Business-process maturity -Data quality and availability -API and integration readiness -Clean-core alignment -Security and authorization design -Joule and SAP Business AI prerequisites -Agent monitoring and governance The goal is to produce a practical roadmap with use cases that can realistically be implemented. ##Custom SAP Agents SAP will deliver many standard Joule agents, but every customer also has unique processes, custom developments and non-SAP applications. Mygo can design and build custom agents that: -Query SAP data using natural language -Execute approved SAP APIs -Call governed function modules and ABAP classes -Coordinate workflows across applications -Analyze operational exceptions -Generate functional and technical documentation -Assist SAP consultants and developers -Automate repeatable business activities ##Joule Studio and BTP Development We help customers build AI-enabled solutions using: -Joule Studio -SAP AI Core -Generative AI Hub -SAP CAP and RAP -SAP Integration Suite -SAP Build -SAP HANA Cloud -Document AI -Event-driven architecture -Fiori and mobile applications Our focus is on integrating AI into the actual business process, not building a separate chatbot that users eventually stop using. ## MCP Servers and Governed SAP Tools Many organizations want to expose SAP capabilities to enterprise AI assistants and agents. Mygo can help build a governed MCP architecture that exposes: -SAP APIs -Function modules -ABAP classes -CDS views -Business objects -Custom applications -External services The framework can include: -SAP user mapping -Authorization checks -Tool allowlists -Input validation -Audit logging -Approval workflows -Human-in-the-loop controls ## Clean-Core and Integration Modernization AI agents need reliable services and clean integration patterns. Mygo can help customers: -Review custom developments -Modernize ECC and S/4HANA integrations -Move suitable extensions to SAP BTP -Build reusable CAP and RAP services -Migrate PI/PO scenarios to Integration Suite -Introduce event-driven integration -Prepare custom code for S/4HANA -Build an AI-ready clean-core architecture ## AI-Enabled SAP Delivery We are also using AI to improve how SAP projects are delivered. This includes: -Solution discovery -Functional-specification generation -Technical-specification generation -Test-case creation -Custom-code analysis -Migration support -Knowledge extraction from past projects -Requirement management -Consultant and developer assistance The objective is not to replace experienced consultants. It is to help them work faster, retain project knowledge and spend more time on decisions that require real business and technical judgment. ##Final Thoughts SAP’s Q2 2026 results show that cloud momentum remains strong. More importantly, they show where SAP believes the market is heading. Cloud ERP, trusted business data, BTP and AI agents are beginning to come together as one connected strategy. Customers should not rush into AI simply because the technology is available. But they should also not wait until every detail is perfect. The right approach is to start with a meaningful business problem, build the required foundations and scale what works. The next phase of SAP transformation will not be defined only by whether a customer has moved to the cloud. It will be defined by how effectively the organization can combine business processes, trusted data, clean-core architecture and governed AI execution.