AI

Before Building Custom AI, Start with SAP-Native Capabilities A practical way to improve enterprise resource planning processes without duplicating technology, data, and cost Artificial intelligence (AI) adoption is now mainstream. McKinsey’s 2025 global survey found that 88% of respondents reported regular AI use in at least one business function. SAP also reported in 2026 that more than 34,000 customers globally already rely on SAP Business AI capabilities embedded across its cloud portfolio. For organizations whose core operations already run on SAP enterprise resource planning (ERP) systems, this raises a practical question: should every process problem lead to a new standalone AI platform, or should teams first examine what is already available in their SAP environment? Start with the Process, Not the AI Product The starting point should be the actual pain points experienced by SAP users, such as: Repetitive manual activities Delayed approvals Difficult information retrieval Time-consuming reporting Limited forecasting and decision support Complex navigation Inconsistent data entry High volumes of employee or supplier enquiries Each pain point should be assessed against the simplest effective intervention before a new AI solution is introduced. The review should progress through the following options: Process simplification Existing SAP standard functionality SAP configuration and workflow automation Better reporting and analytics User guidance and training Embedded SAP AI and Joule capabilities A governed custom AI extension Why SAP-Native AI Deserves First Consideration SAP Business AI is designed to bring AI into established business processes using SAP data, process context, roles, and workflows. It includes embedded AI use cases, Joule assistants and agents, and tools for developing custom extensions through SAP Business Technology Platform (SAP BTP). Starting with SAP-native capabilities can provide several advantages: Reduced integration and implementation complexity Better alignment with existing SAP processes and controls Reuse of established roles and authorizations Less duplication of enterprise data Simpler support and lifecycle management Greater value from existing SAP investments For common ERP process improvements, this route may be more efficient than building and operating a separate AI platform. However, the business case must still be validated. Licensing, implementation effort, integration, data readiness, security, and total cost of ownership (TCO) should be assessed for the specific SAP product, deployment model, release level, configuration, and commercial arrangement. Do Not Use AI to Compensate for Weak Foundations AI should not be used to mask weak process design, poor data quality, incomplete configuration, unnecessary customization, or insufficient user training. Automating a broken process usually makes the problem faster and harder to govern. Before approving an AI initiative, business and technology teams should confirm: The business problem and desired outcome are clearly defined. The underlying process has been simplified where possible. Required data is accurate, accessible, and appropriately governed. Existing SAP capabilities have been evaluated rather than assumed inadequate. Benefits can be measured against cost, risk, and operational impact. When Custom or Third-Party AI Is the Right Choice Custom or third-party AI may still be justified when: SAP does not provide the required capability. The use case spans several enterprise platforms. The organization has unique business or industry requirements. Greater control over models, data, or deployment is required. SAP capabilities do not meet language, compliance, or usability needs. The SAP commercial model is not viable. The capability provides genuine strategic differentiation. Where customization is required, it should preferably be delivered as a governed extension that protects the ERP clean core, rather than as a complex modification to the underlying SAP system. A Practical Decision Principle Use SAP-native AI where it adequately improves an existing SAP process. Build custom AI only where there is a clear functional, strategic, technical, or commercial justification. This principle connects AI investment to real business pain points, protects existing technology investments, and reduces the risk of disconnected solutions that duplicate capabilities already available within the SAP ecosystem. Sources and Further Reading McKinsey & Company, “The State of AI: Global Survey 2025” SAP, “SAP Labs to Hire and Train AI Scientists and Machine Learning Engineers in Partnership with IMDA” SAP Business AI and Joule SAP Learning, “Explore How Customers Benefit from Out-of-the-Box and Custom AI” SAP Learning, “Building Custom SAP Business AI Solutions” Conclusion: Treat AI as Process Improvement, Not Platform Acquisition An SAP-native-first approach does not mean “SAP only.” It means using the most efficient and governable layer first, then escalating to custom AI when the business case is clear. This protects the clean core, reduces duplication, and directs investment toward measurable operational outcomes rather than technology for its own sake. Ready to identify practical AI opportunities in your ERP landscape? Weigh your pain points against SAP-native capabilities first, and build custom only where there’s a clear justification. SOLGulf works with enterprise SAP teams on exactly this kind of assessment — talk to us to get started. About the Author Bilal Ahmed Principal Enterprise Architect, SOLGulf Bilal Ahmed is a Principal Enterprise Architect with more than 20 years of experience across enterprise architecture, digital transformation, cloud, and solution delivery. His work focuses on translating business priorities into governed technology decisions, particularly for government and regulated environments.

5 Ways Project Managers Can Lead in Saudi Arabia’s Year of AI Artificial Intelligence (AI) is no longer a side experiment. In 2026, the Saudi Cabinet declared the Kingdom’s Year of Artificial Intelligence, placing AI at the heart of Vision 2030, Saudi Arabia’s national transformation program. For project managers (PMs), that shift raises an urgent question: as AI changes how work gets delivered, will you lead the change or simply react to it? The pressure is real, and the numbers show why. According to Gartner, by 2030 around 80% of routine project management tasks, such as data collection, tracking, and reporting, will be handled by AI. But automation is not the same as replacement. In a 2025 Gartner survey of more than 700 chief information officers (CIOs), leaders expected that by 2030, 75% of technology work will be done by people working alongside AI, and only 25% by AI on its own. The message for PMs is clear. AI will take on more of the busywork, but leading the outcome stays firmly human. Here are five practical ways to step into that role. 1. Start with the data, not the tool The most useful mindset shift is simple: AI projects are not software projects, they are data projects. The Project Management Institute (PMI) has made this point for years, and teams learn it the hard way when a promising model fails because the data feeding it was messy or incomplete. Before choosing a tool, get clear on what data you have, what shape it is in, and whether it can actually support the outcome you want. 2. Close the gap between the demo and the deployment Most AI initiatives don’t fail because the model was weak. They stall in the gap between an impressive demo and a working deployment, where scope drifts, data isn’t ready, and stakeholders never quite align. That gap is exactly where good project managers earn their keep, by turning a promising prototype into something that ships and holds up in the real world. 3. Scope every initiative around a measurable outcome AI is easy to get excited about and easy to waste money on. The fix is discipline: tie every initiative to a clear, measurable business outcome before the build begins. This is where a strong Project Management Office (PMO), the team that sets delivery standards across an organization, adds real value, by keeping projects honest about the results they promised. 4. Double down on the skills AI can’t replace Here is what surprises people: the difference between a PM getting a little value from AI and one getting enormous value is rarely the tool. It’s communication, critical thinking, collaboration, and creativity. As routine tasks get automated, these human skills become the real differentiator, because AI rewards the leaders who ask sharper questions and frame better problems. 5. Stay accountable for what ships When AI drafts the plan or flags the risk, someone still has to own the decision. Keeping a human clearly accountable isn’t a brake on speed, it’s what lets you ship with confidence and keep the trust of clients and stakeholders. The best PMs treat accountability as part of good delivery, not paperwork added at the end. The bottom line The Year of AI won’t be won by the teams that adopt the most tools. It will be won by the people who can turn AI into real, delivered results, on time and on outcome. This is exactly where SOLGulf’s Agentic AI Delivery & Operations practice helps: we move organizations from AI readiness to production, bringing the delivery discipline that makes results stick. If you’re ready to turn AI ambition into working outcomes, reach out to the SOLGulf team, and we’ll show you what it looks like in practice. About the Author Syed Tufail Ahmed,AI & Digital Program Manager, SOLGulf Syed Tufail Ahmed is an AI governance strategist and program leader with around 25 years of enterprise experience across digital banking, healthcare analytics, and technology. At SOLGulf, he serves as AI & Digital Program Manager and heads the Project Management Centre of Excellence (PMCoE) for a Saudi government client, leading the delivery of AI and digital transformation initiatives. He is the author of Human in the Loop: Reclaiming Human Authority in an Age of Intelligent Systems (2026) and a Global Ambassador for the Global Council for Responsible AI (GCRAI). His work centers on helping organizations adopt AI in ways that scale human judgment rather than replace it. He has been recognized by Thinkers360 among the top global voices in AI governance, ethics, and safety, and holds a Certified Expert credential in AI Governance. Syed writes and speaks regularly on responsible AI, project delivery, and the evolving role of the project manager. He holds a Bachelor of Engineering in Instrumentation and Electronics and is based in Riyadh, Saudi Arabia. Connect with him on LinkedIn: linkedin.com/in/tufailsa