Better coding with AI

Image: SAP / Getty Images

Der Artikel kann nur mit aktiviertem JavaScript dargestellt werden. Bitte aktiviere JavaScript in deinem Browser und lade die Seite neu.

Generative AI has tremendous potential for development teams. It makes the programming process more efficient by generating high-quality code and improving collaboration.

by Michael Ameling, Senior Vice President, Head of Intelligent Enterprise & BTP Foundation SAP

Programming without in-depth programming knowledge? Thanks to generative artificial intelligence (AI), you can. Even if you already have the necessary skills, the technology offers tremendous potential. Generative AI makes software development faster, more efficient, and more productive, and it will have a lasting impact on programming.

SAP has clearly defined goals in this field: First, every developer in the SAP environment, including partners and customers, should be able to use generative AI and become an “AI developer.” To achieve this, SAP has developed new tools that support the development and deployment of intelligent data applications at scale.

SAP is also simplifying the interaction with large language models (LLMs) so that developers can create intelligent applications. With the addition of a vector engine to the SAP HANA Cloud database capabilities, developers can add context to largely generic LLMs with company-specific and current data. This increases the relevance of AI-generated results.

Finally, the SAP Business Technology Platform (SAP BTP) is the first choice for adding AI to applications. Generative AI is embedded in SAP cloud solutions through the SAP BTP. This means that SAP customers’ and partners’ development teams always have access to the latest AI technologies and can easily and securely build business applications on SAP BTP.

Faster application development

How can development teams actually use generative AI with SAP BTP? One answer is SAP Build Code. This offering within SAP BTP, designed specifically for software developers, includes AI-enabled tools and services for working with Java and JavaScript. This includes Joule, the new digital assistant recently introduced by SAP.

With Joule, development teams can automate complex tasks directly in the programming environment. They can use speech to generate application logic, data models, and test scripts, and integrate newly developed applications into the platform.

SAP Build Code also promotes collaboration between professional developers and other technically skilled employees from different parts of the organization in so-called fusion teams. With SAP Build Code, SAP bundles the familiar tools of both groups to achieve results in less time. This ensures seamless integration with SAP application systems.

Vector Database Supercharges AI Models

Keep in mind that AI can only do what it’s trained to do. If the underlying data is outdated or does not relate to specific operational issues, it can lead to inadequate results or even “AI hallucinations” — objectively incorrect results.

This is why new AI applications rely on “vector embeddings” to map complex and diverse data into a relatively low-dimensional vector form. Simply put, embeddings allow complex and unstructured data to be efficiently represented in a structured form. This is critical for high-quality AI results.

Whatever AI solution you end up with: It is critical that the solution is not simply an additional module added to an existing application. At SAP, it is fully integrated into the SAP cloud solutions, offering users the greatest possible potential.


Leave a Reply

Your email address will not be published. Required fields are marked *