SAP in the AI Race: Catch-up, Ambitions, and the Reality of the Enterprise Giant
The current wave of generative artificial intelligence (AI) is reshaping industries and putting massive pressure on enterprise software titans like SAP to deliver not just promises, but tangible, business-altering AI solutions. SAP, a company whose software underpins a significant portion of global commerce, finds itself in a high-stakes race to redefine its value proposition in an AI-centric world. This isn’t just about adding AI features; it’s about embedding intelligence into the core of business operations – an area SAP has historically dominated but where it now faces the perception of having fallen behind more agile, cloud-native competitors. For SAP, this is more than just a trend; it’s an existential necessity to maintain market leadership as AI fundamentally changes the optimization and management of business processes, SAP’s traditional strength.
This article critically examines whether SAP is actually closing the perceived AI gap. What concrete progress underpins the “Business AI” strategy, particularly as presented at the recent Sapphire conference in 2025? How crucial is the strategic partnership with Microsoft, and what are its real-world implications for customers? And beyond the hype: what are the true opportunities, risks, and maturity level of SAP’s AI offerings in the demanding enterprise landscape? The success or failure of SAP’s AI strategy will have far-reaching consequences for thousands of companies worldwide, influencing their own digital transformation roadmaps and competitiveness.
Context and Background: SAP’s Long Road to Business AI
For decades, SAP has established itself as the global standard for Enterprise Resource Planning (ERP) software, based on the integration of business processes and real-time data processing. This history provides the company with a potential goldmine for AI: a vast repository of business data and process knowledge. Early forays into AI-related technologies manifested in SAP Leonardo, which, as a “digital innovation system,” combined IoT, machine learning (ML), Big Data, and analytics on the SAP Cloud Platform. Although ambitious, Leonardo was sometimes perceived more as a collection of services than a deeply integrated AI strategy, which later led to its reorganization into SAP AI Business Services and a more embedded approach. The development towards S/4HANA and the Business Technology Platform (BTP) laid the foundation for modernizing the SAP landscape and enabling new technologies like AI.
In today’s enterprise software market, particularly in ERP, CRM, SCM, and HCM, AI is becoming a crucial differentiator. Here, SAP faces strong competitors: Salesforce with its Einstein AI in the CRM space, often perceived as more agile; Oracle, which is advancing embedded AI in its Fusion Cloud Applications and often accuses SAP of a fragmented approach and additional AI costs; and Microsoft, a significant partner with Azure AI Services but also a competitor with Dynamics 365. Added to this are hyperscalers like AWS and Google Cloud, which offer foundational AI services and are also partnered with SAP.
This constellation has contributed to the perception of an “AI lag” at SAP. Historically, SAP was not always seen as an AI pioneer; analysts and customers perceived its AI efforts as maturing more slowly or being less coherently integrated compared to cloud-native competitors. Criticisms sometimes included the complexity of AI integration into highly customized legacy SAP systems and the challenge of moving from isolated AI projects to enterprise-wide AI adoption. Gartner Peer Insights for data science and machine learning platforms show a competitive field, and the Forrester Wave for CRM software 2025 classifies SAP as a “Contender,” with strengths in back-office integration but less pronounced AI functions in the CRM area compared to the “Leaders.”
SAP’s deep ERP dominance proved to be a double-edged sword for AI adoption. While the immense data and process knowledge is an invaluable fuel for AI, the complexity and customizations built up over decades in legacy systems – often referred to as technical debt – slowed down rapid, broad AI integration and contributed to the perception of a lag. This inherent inertia contrasts with the agility of cloud-native competitors who could develop AI from scratch. The strategic shift from SAP Leonardo, a rather standalone innovation brand, to “Business AI,” an embedded capability, reflects a crucial realization: for AI to succeed at SAP, it must be inextricably linked to core business processes and applications, not an add-on. “Business AI” aims to make AI an intrinsic part of the SAP solutions already in use, which, if well executed, lowers adoption hurdles. The competitive landscape is also complex: SAP uses hyperscalers for infrastructure and AI services but simultaneously competes with their application offerings. This “coopetition” shapes SAP’s partner strategy and market positioning, where the focus must be on creating unique business process intelligence on these platforms.
Current Status and Innovations: SAP’s Offensive with Business AI
SAP is now aggressively advancing its “Business AI” strategy, emphasizing AI that is relevant (solves real business problems), reliable (precise and consistent), and responsible (ethical and compliant). This represents a clear strategic shift to build trust and differentiate from generic AI. The core idea is the direct embedding of AI into business processes across the entire cloud portfolio, aiming to achieve 400 embedded AI use cases by the end of 2025 – an increase from the over 200 currently available. CEO Christian Klein emphasized at Sapphire 2025 that the goal is to deliver tangible AI value now or within the next six months, moving from long-term promises to concrete implementation.
Joule: The Omnipresent AI Copilot
Joule, SAP’s flagship AI copilot, is evolving into an “omnipresent” assistant, embedded not only in SAP applications but also extending to third-party systems. Key enhancements announced at Sapphire 2025 include:
- Proactive Support & Action Bar: Based on the technology from the WalkMe acquisition, Joule will feature an “Action Bar” (generally available Q3 2025) that analyzes user behavior and proactively offers insights, recommendations, and task automation. This transforms Joule from a reactive to an anticipatory assistant.
- Joule Agents: An expanded network of pre-built and custom AI agents is intended to autonomously (with human oversight) automate and orchestrate complex end-to-end business processes in areas such as finance, supply chain, and human resources. Examples include agents for quote creation, expense validation, and manufacturing monitoring.
- Joule Studio: As part of SAP Build, it offers low-code/no-code capabilities for customers and partners to create custom skills (Q3 2025) and AI agents (Q4 2025) for Joule.
- AI Agent Hub & LeanIX Integration: A new hub (generally available Q4 2025), integrated with SAP LeanIX, is intended to provide centralized inventory, governance, and mapping of AI agents to business processes.
SAP AI Foundation: The “Operating System for Business AI”
Introduced at Sapphire 2025, the AI Foundation is positioned as a unified operating system for developers to build, extend, and run custom AI solutions on the SAP BTP. Components include an AI Agent Runtime, a Prompt Optimizer (co-developed with Not Diamond to automate the creation of effective prompts), Joule Studio, AI Evaluation Services, and enterprise-wide Knowledge Graphs connected to Datasphere (Q3 2025). The goal is to accelerate AI innovation and simplify AI operations for businesses.
SAP Business Data Cloud: Fuel for Intelligent Applications
The SAP Business Data Cloud is evolving beyond pure analytics to become the foundation for intelligent applications by unifying SAP and non-SAP data in a single semantic layer while preserving business context. SAP Datasphere forms an important basis for this. New “intelligent applications” such as People Intelligence (HR insights from SuccessFactors), Spend Control Tower, and Green Ledger are being introduced. Partnerships with Databricks (native embedding, zero-copy data sharing), Palantir (connectivity with Foundry), and Google BigQuery are crucial for this strategy.
The Microsoft Partnership: Deepened Relations for AI Progress
A key announcement is the bidirectional integration of Joule and Microsoft 365 Copilot, enabling employees to access information and complete tasks across both environments, aiming for a seamless, unified AI experience (general availability for some aspects expected Q3 2025, Joule Custom Engine Agent in the Agent Store by end of Q2 2025). This directly addresses friction in the user experience and aims to eliminate vendor silos. Furthermore, SAP BTP Services will be available on the Azure Marketplace (USA H2 2025), and SAP Business Data Cloud as well as SAP Databricks on Azure are planned for Q3 2025. Azure is also the first provider to enable SAP’s 99.95% SLA option for SAP Cloud ERP Private. The partnership also extends to co-innovation, security enhancements, and programs to accelerate RISE with SAP on Azure.
Other Key Partnerships and Developer Enhancements
Other important collaborations include Perplexity AI for integrating real-time web insights into Joule, AWS as part of an AI Co-Innovation Program to create generative AI solutions with Amazon Bedrock and SAP BTP, Google Cloud for BigQuery integration and the Agent2Agent protocol, and NVIDIA for reasoning models in Joule Agents and AI in robotics. For developers, there are AI-powered ABAP tools for ECC migration, integration of custom LLMs into the Integration Suite, and marketplace availability via AWS and GCP.
The multitude of these announcements underscores SAP’s determination not only to catch up in the AI field but to take a leading role. The following table summarizes the most important AI-related announcements from SAP, particularly from Sapphire 2025:
Table 1: SAP’s Key AI Announcements (Sapphire 2025 & Recent Developments)
| Announcement/Initiative | Kurzbeschreibung & Strategisches Ziel | Schlüsselfunktionen/Auswirkungen | Angegebene Verfügbarkeit |
| Joule Omnipräsenz & Action Bar | Joule als proaktiver, anwendungsübergreifender Assistent | WalkMe-Technologie, proaktive Empfehlungen | GA Q3 2025 |
| Joule Agents Erweiterung | Automatisierung komplexer End-to-End-Prozesse | Vorgefertigte Agenten (Finanzen, SCM etc.), Agent2Agent Protokoll (Google) | Viele Q2/Q4 2025 |
| Joule Studio | Ermöglichung der Low-Code-Erstellung benutzerdefinierter KI-Agenten | Skill Builder, AI Agent Builder in SAP Build | Q3/Q4 2025 |
| AI Agent Hub | Zentralisierte Governance und Verwaltung von KI-Agenten | Integration mit SAP LeanIX, Mapping zu Geschäftsprozessen | Q4 2025 |
| SAP AI Foundation | Einheitliches Betriebssystem für die Entwicklung von Unternehmens-KI | AI Agent Runtime, Prompt Optimizer, Joule Studio, Knowledge Graphs | Teile GA, Rest 2025 |
| Prompt Optimizer | Automatisierte Erstellung optimierter Prompts für verschiedene LLMs | Co-Design mit Not Diamond, Reduktion manuellen Engineerings | Teil der AI Foundation |
| Business Data Cloud Erweiterungen | Intelligente Anwendungen auf Basis einer einheitlichen Datengrundlage | People Intelligence App, Integration mit Databricks, Palantir, Google BigQuery | Teile GA, Rest H2 2025 |
| Perplexity Partnerschaft | Integration von Echtzeit-Web-Erkenntnissen in Joule | Verifizierte öffentliche Daten zur Anreicherung von Unternehmensantworten | Integriert in Joule |
| Microsoft Copilot Integration | Nahtloser Datenzugriff und Aufgabenerledigung zwischen SAP und Microsoft-Umgebungen | Bidirektionale Integration von Joule und M365 Copilot | Aspekte Q2/Q3 2025 |
| AWS Co-Innovation Program | Unterstützung von Partnern bei der Entwicklung von GenAI-Lösungen mit Amazon Bedrock und SAP BTP | Technische Ressourcen, Cloud-Credits, Expertise | Gestartet Mai 2025 |
| Databricks on BDC | Native Einbettung von Databricks in die Business Data Cloud | Zero-Copy Data Sharing, Nutzung von Databricks-Fähigkeiten mit SAP-Daten | Q2 2025 (Partner Connect) |
| Palantir for BDC | Nahtlose Konnektivität zwischen Palantir Foundry/AIP und SAP Business Data Cloud | Harmonisierte Datengrundlage für Kunden | Angekündigt Mai 2025 |
SAP’s current AI strategy is heavily dependent on a broad ecosystem of partnerships. This signals the recognition that innovations in AI cannot occur in isolation and that the integration of best-of-breed technologies is necessary – a remarkable shift from a historically more proprietary stance. The Business Data Cloud, AI Foundation, and Joule form a networked value chain: The Business Data Cloud delivers the contextualized data, the AI Foundation provides the tools for creating AI models and agents, and Joule offers the interface for delivering AI-driven insights and actions to users. Joule’s success thus directly depends on the strength of the underlying data and AI platforms, because AI, as is often said, is “only as good as the underlying data.”
The vision of an “omnipresent” Joule and statements like “AI will become the new user interface” indicate the long-term ambition to fundamentally change how users interact with enterprise software, potentially abstracting much of the traditional application complexity. This is a transformative, but also extremely challenging goal. Simultaneously, the introduction of Joule Studio for low-code/no-code agent development aims to democratize AI development, while the AI Agent Hub in LeanIX and the focus on AI ethics address the critical need for governance and control as AI becomes more widespread. This balance between creation and control will be crucial for enterprise adoption.
Critical Analysis: Opportunities, Risks, and Maturity of SAP’s Business AI
SAP’s ambitious AI strategy holds significant opportunities but is also associated with considerable risks and challenges regarding market readiness.
Opportunities: The Potential of SAP’s Business AI
SAP possesses a core strength that can be invaluable in the AI era: a deep understanding of end-to-end business processes across diverse industries and access to vast amounts of transactional and operational data within its customers’ systems. If this data and process context knowledge can be effectively leveraged for “Business AI,” it can yield highly relevant and directly actionable insights, unlike generic AI models. The strategy of embedding AI directly into core applications like S/4HANA, SuccessFactors, or Ariba and their workflows promises AI support exactly where users work, reducing friction and promoting adoption. SAP itself speaks of delivering “AI that actually understands all your business processes and data.”
The company forecasts productivity gains of up to 30% through Business AI, and specific use cases already show significant time savings, such as 90% faster category management processes or 90% faster root cause analysis at AMD. The deepened partnership with Microsoft, particularly the integration of Joule and Microsoft 365 Copilot, could act as a force multiplier, creating a unified digital work environment that overcomes data silos between SAP’s transactional world and Microsoft’s productivity suite. The focus on “Business AI” to solve concrete business problems like demand planning, predictive maintenance, or fraud detection can also deliver measurable value more quickly. Lastly, SAP’s emphasis on responsible AI, transparency, and data protection, including recognition by the World Benchmarking Alliance, can be an important differentiator for companies aware of AI’s risks.
Risks: Hurdles on the Path to AI-Powered Enterprise
Despite promising opportunities, SAP and its customers face significant challenges. Implementing the ambitious roadmap with numerous innovations around Joule, new agents, and the AI Foundation is a massive task where timelines and quality must be met. Historically, SAP has occasionally been criticized for the pace of innovation delivery.
Customer acceptance depends on several factors. A key point is the need for a “Clean Core”: the full benefits of new AI functions, especially those embedded in S/4HANA Cloud, often require standardized processes and minimal customizations. However, many long-standing SAP customers operate highly customized legacy systems, making the transition complex and costly. Existing technical debt in customer landscapes can significantly hinder the introduction and effectiveness of new AI solutions, as AI does not magically fix underlying system problems. The cost and ROI justification for premium AI functions and consumption-based services will also be a hurdle; Oracle, for example, criticizes SAP for additional AI costs. Furthermore, companies often lack internal capabilities to implement and use advanced AI solutions and face the complexity of integrating new AI tools into existing, often heterogeneous IT landscapes.
SAP must also hold its own against competitors who are also innovating rapidly. The principle “no context, no value” means that AI results heavily depend on the quality and governance of the underlying business data – an area where many organizations struggle with silos and quality issues. Finally, there is always the risk of “agent washing” or hype if the initial capabilities of AI agents do not match the grand vision of autonomous multi-agent systems.
Maturity: Between Vision and Market Reality
The vision of Joule as an omnipresent assistant with proactive capabilities and a network of specialized agents is forward-looking. While basic Joule functions are available, many advanced agentic functionalities and the full “omnipresent” experience (e.g., Action Bar via WalkMe) are planned for Q3/Q4 2025 or later. The market readiness of truly autonomous, cross-system AI agents for complex tasks is still developing industry-wide. The SAP AI Foundation, as a newly introduced “operating system for Business AI,” must first prove its maturity through rapid adoption by developers and partners, as well as the robustness of its components like the Prompt Optimizer and Knowledge Graph.
SAP points to over 200 existing embedded AI use cases, indicating a certain level of maturity in specific areas (e.g., AI in S/4HANA for invoice matching, demand planning). However, the transformative potential often lies in the newer generative AI and agentic capabilities. The announcements at Sapphire 2025 generated considerable buzz, and analysts like those at ARC Advisory Group attested to a “bold, coherent, and surprisingly open vision”, but also warned that execution is crucial and the path complex. The claim of “30% productivity gains” is ambitious and requires substantiation in diverse customer environments.
In comparison, Salesforce Einstein is considered mature in the CRM area for predictive insights and automation. Oracle emphasizes its embedded AI in Fusion Cloud Apps and a unified data model, which is intended to enable faster innovation cycles. Microsoft’s AI capabilities via Azure AI and Copilot are broad and evolving rapidly. SAP’s unique selling proposition is its deep integration with its own business processes and data. The maturity question concerns less generic AI capabilities and more the effective implementation of these into tangible business results within the complex SAP ecosystem. IDC notes that SAP’s strategy combines AI, applications, and data to improve functionality and efficiency, but success depends on vendor expertise and system design. Forrester positions SAP CRM as a “Contender,” suggesting lower maturity in AI-driven CRM functions compared to “Leaders.” Constellation Research highlights Joule as key but emphasizes that AI is only as good as the data.
The necessity of a “Clean Core” is not just a recommendation but a fundamental dependency for unlocking much of SAP’s advanced AI. Legacy customizations and associated technical debt directly block the path to utilizing these new AI capabilities. This creates a multi-year transformation journey for many customers before they can fully benefit, potentially widening the gap to more agile competitors in the short term. The “time-to-value” for SAP’s AI is thus inextricably linked to the customer’s “time-to-clean-core.”
There is a tension between SAP’s messages of speed and immediate value (“AI now,” rapid use-case implementation) and the reality of AI adoption in enterprises, which is inherently complex, often lengthy, and involves data readiness, integration, change management, and governance. This could lead to a discrepancy between SAP’s marketing and the customers’ operational reality. However, SAP’s strong focus on “Responsible AI” could become an increasingly important competitive advantage as companies become more aware and cautious of the ethical risks, bias, and compliance issues associated with AI. This appeals to the governance-minded nature of many large SAP customers. A subtle but profound implication of the AI strategy is the mention of a potential shift towards an “Outcome-as-a-Service” model. If AI can reliably automate and optimize processes to deliver guaranteed results, the traditional software licensing model could evolve – a long-term vision showing where deep AI integration might lead.
The following table contrasts SAP’s AI-driven opportunities with implementation realities:
Table 2: SAP Business AI – Opportunities vs. Implementation Realities
| SAPs KI-gestützte Chance/Versprechen | Schlüsselaspekte der Kundenimplementierung/Herausforderungen |
| Joule als omnipräsenter proaktiver Assistent | Erfordert „Clean Core“ & S/4HANA Cloud Adaption; Nutzerakzeptanz und Training |
| Autonome Joule Agents zur Prozessorchestrierung | Signifikanter Aufwand für Datenqualität, Governance und Integration; Komplexität der Entwicklung/Verwaltung benutzerdefinierter Agenten, potenzielle Skill-Gaps |
| AI Foundation für benutzerdefinierte Unternehmens-KI | Bedarf an Entwickler-Know-how für BTP und AI Foundation; Sicherstellung von Skalierbarkeit und Wartbarkeit |
| Business Data Cloud für einheitliche Einblicke | Notwendigkeit einer robusten Datenstrategie, Verbindung disparater Quellen; semantische Harmonisierung |
| 30 % Produktivitätssteigerung durch Business AI | Abhängig von erfolgreichem Change Management, Prozessneugestaltung und Überwindung technischer Schulden; ROI-Nachweis für spezifische Anwendungsfälle |
Outlook and Assessment: SAP’s AI Future Between Aspiration and Reality
SAP’s AI journey is critically important for virtually every large enterprise using SAP. The company is undertaking a determined, comprehensive, and increasingly open effort to place AI at the center of its offerings. The next steps include the continuous introduction and refinement of Joule, Joule Agents, and the AI Foundation components throughout 2025 and beyond, with adoption rates and real-world performance needing close observation. The expansion of embedded AI use cases across the SAP cloud portfolio towards the goal of over 400, as well as the deepening of strategic partnerships, especially with Microsoft, but also with other AI innovators and data platform providers, will be in focus. The interoperability of AI agents, for example, with Google Cloud’s Agent2Agent protocol, will be critical. Another focus will be on supporting customer migration to S/4HANA Cloud and achieving a “Clean Core” as a prerequisite for full AI utilization. The development of industry-specific AI solutions based on the Business Data Cloud and the AI Foundation, as well as the evolution of AI pricing models, will also be important developments.
Even if the perception of an “AI lag” may have been justified in the past, SAP’s recent announcements, particularly around Joule, Business AI, and strategic partnerships, demonstrate significant ambition and concrete steps to catch up and even leapfrog in areas where business process expertise offers an advantage. However, the path is complex. The hype around AI must be tempered by the reality of enterprise transformation. The responsibility lies not only with SAP to deliver innovative technologies but also with customers to undertake the necessary modernization (Clean Core, Data Governance) to utilize them. The partnership with Microsoft is an important accelerator, especially for user experience and data integration. However, SAP’s success will ultimately depend on its ability to deliver uniquely valuable “Business AI” that competitors cannot easily replicate.
SAP’s AI transformation and its customers’ digital transformation are inextricably linked. SAP needs modernized customers to adopt its AI; customers need SAP’s AI to modernize their core business processes. This symbiotic relationship will determine the pace and success of SAP’s AI strategy. The ultimate battlefield for SAP is not just adding AI features, but ensuring its platform remains the “intelligent core” of the enterprise. This means AI that is not only powerful but deeply understands and orchestrates complex, cross-functional business processes in a way that standalone AI platforms or more narrowly focused competitors cannot.
Companies should approach SAP’s AI offerings with informed optimism: engage with the new tools, pilot relevant use cases, but also critically assess the prerequisites and invest in the foundational changes necessary for long-term AI success. The promise is substantial, but so is the required commitment. While SAP’s AI ambitions are vast and recent progress is noteworthy, the road ahead is fraught with implementation risks and customer adoption challenges. The narrative is shifting from “Does SAP do AI?” to “How effectively can SAP and its customers implement and scale Business AI for tangible, transformative results?” The industry will be watching not just for new features, but for impactful case studies and proven ROI.
