In the age of AI, the skills we build today will define how we work tomorrow. That’s why we’ve launched a monthly blog series that spotlights the latest AI learning assets and upcoming learning opportunities.

Building on my first blog, where I introduced you to our AI live sessions Beyond Hype (which are all about getting productive with SAP Business AI), I wanted to address some of your most asked questions. But before I do that, I’m excited to share that the sessions are now ungated and no longer require a SAP Learning Hub subscription!

1. How does SAP ensure that frontier models can be trusted?
   It’s a shared responsibility across SAP, the model vendor, and the customer, managed through a layered and continuous improvement approach:

• SAP: Selects enterprise-grade vendors; conducts security and responsible AI reviews; integrates models behind SAP identity, authorization, logging, and data protection; applies guardrails; monitors behavior and coordinates fixes/updates.
• Model vendor: Owns model quality and defects; contractually obligated to address safety issues and deliver improvements when identified by either party.
• Customer: Configures authorizations and data access; defines permissible use cases and safety settings; participates in monitoring and feedback.SAP cannot change a vendor’s model weights or eliminate all model-intrinsic limitations. Instead, SAP mitigates risk through configuration and guardrails, protects business data via SAP’s application security, monitors performance, and escalates issues to the vendor under defined contracts and SLAs. If needed, SAP can adjust configurations, deploy compensating controls, or replace/disable a model to protect customers.Find more answers in the publicly available SAP Business AI – Responsible AI FAQ.

1. What are the most common security mistakes SAP customers make when implementing an AI?For AI features embedded in existing business applications, the underlying system typically enforces data controls and authorization; the AI component itself doesn’t determine access, the host application does. When teams build custom AI applications, or add new components, they must implement the right authorization and access controls for those pieces. Overall, AI adoption is still experimental. While custom implementations vary widely by organization, one recurring theme is content moderation: Some teams keep it on, while others relax or disable it to support domain-specific use cases (for example, medical contexts). That configuration choice is worth highlighting. Beyond that, patterns are highly situational and we refrain from framing them as mistakes so much as considerations that differ by approach.Watch AI Security – What is it about? And understand how SAP is developing, deploying, using, and selling AI systems with the highest ethical, security, and privacy standards. 

1. Why should I use SAP AI Core for agent assisted coding?
   
   AI Core uses a secure environment where the LLMs are hosted, and the data is not used for further training compared to public LLMs. To deliver consistent, secure, and compliant outcomes across your landscape, you can bring the AI coding environment of your choice (e.g. Claude Code or Cline).You can learn more about agent assisted coding with Cline and AI Core in the SAP Architecture Center.

Agent‑assisted coding began as a developer aid to speed up prototyping, refactoring, test generation, and debugging. It has since evolved into a broader set of AI agents that translate natural language into software artifacts, connect design and code, scaffold data and infrastructure, and automate tedious workflows. This shift lowers the barrier to software creation and lets more roles contribute directly, effectively democratizing development by allowing anyone to create natural language prompts instead of traditional codes. These developments unlock a wide range of new use cases, enabling you to work more creatively and productively while creating higher-quality technical documentation.

One of these new use cases is spreadsheet automation, which is especially valuable for non-technical users. Imagine you are working with a set of marketing data spread across different channels. Instead of manually cleaning and analyzing it, you simply ask the system to help you create a structured spreadsheet with performance insights. What happens next is that the system automatically organizes your data and generates a clear Excel file. It can include multiple sheets such as campaign performance, channel performance, and daily trends. At the end, it also adds a short summary that highlights the key insights from your data. Behind the scenes, the system generates a custom Python script that processes your raw data and builds the Excel file automatically. The important part is that you don’t need to write any code or manually structure the spreadsheet, you can simply review the output and use it directly.

Another example comes from a product management use case, where you might want to build a new time tracking feature. You can start by asking the system to support you with competitive research. For example, you explore how existing tools like Asana and ClickUp handle time tracking and what features they offer. The system gathers and summarizes this information into a structured overview of competitors and their capabilities. From there, you can take the next step and ask it to generate a product requirements document (PRD). Instead of starting from a blank page, you receive a complete draft that includes goals, success metrics, and a clear breakdown of requirements such as “must-have”, “should-have”, and “nice-to-have” features, as well as non-functional requirements.

But it doesn’t stop there: As a product manager, you can go further and ask the system to break the PRD down into actionable work items. This means it can generate epics and tasks that you can directly use in your planning tools. As a result, you can review, adjust, and refine the output together with your team. Rather than manually compiling research, writing documents, and structuring tasks from scratch, you start from a strong, AI-generated foundation that accelerates your entire product development process.

Now it’s your turn! Try yourself with this sample scenario and use Cline to rapidly build enterprise-grade SAP CAP Application with Fiori UI.

Want to dive deeper into the content? Rewatch the session and access the session material:

• All Things Business AI: Latest Insights, Updates & Live Demos (recording) 
• Build AI securely – Avoid Pitfalls in the Development and Operations Lifecycle (recording)
• Agent Assisted Coding with AI Core (recording)

Hi Joule, where should I get started with the latest learning resources? *

Apart from live enablement sessions, explore our latest on-demand learning resources:

• NEW! Driving SAP Joule for Consultants Adoption Learning Journey: Understand SAP Joule for Consultants value across the project lifecycle, enable consultant daily use, engage stakeholders, deploy SAP’s adoption asset library, and use usage data to optimize and demonstrate business impact.
• NEW! Building Confidence in AI with SAP Learning Journey: Understand Responsible AI in SAP; how ethics, security, privacy, and compliance govern data, how SAP safeguards via policies and certifications, and how customers control usage and involve experts to drive trusted, confident adoption.
• NEW! SAP's AI Golden Path: The SAP's AI Golden Path is the starting point for developing AI applications across the SAP ecosystem. It contains recommendations, best practices, and tutorials to help you understand the AI technology stack, identify suitable tools and services, and design, deliver, and extend enterprise-grade AI solutions on SAP technology.

* Powered by the machine learning technology of SAP AI Business Services, personalized recommendations are generated for employees, in one place on the Learning home page, based on employee data.

As we look ahead, we’re excited to share with you a new lineup of learning opportunities designed to help you expand your knowledge and make AI an advantage in your role:

Stay tuned for more, and happy learning! 🚀

Further Resources:

SAP AI Core Security Artificial Intelligence Joule