Will developers be replaced by the artificial intelligence soon?
While this question sparks a lot of debate, AI is already transforming the way developers build, maintain, and optimize integrations. There’s a lot of hype around AI these days – everyone wants it, and SAP even claims to have AI for the SAP Integration Suite. But let’s take a closer look at what’s actually possible in practice. To explore this, Daniel Graversen from Figaf invited Radosław Ruciński from Sygeon to a webinar. Together, they discussed the current capabilities of AI, its realistic applications in SAP Integration development, and how it may shape the work of developers in the near future.
AI-Assisted iFlow Generation
One of the first topics was using AI to generate iFlows from text prompts. While SAP provides tools for this, access is still limited and they can be used for rather simple flows – which developers can build manually in minutes. It is questionable if AI provides real help for developers in this case..
Radek demonstrated a prototype that could create an iFlow from a description, showing that AI can handle simple flows, but may struggle with more complex scenarios. The discussion highlighted that the real potential of AI lies in empowering business users to automate basic processes, allowing developers to focus on more advanced tasks.
AI for Mapping and Data Transformation
Another key topic was AI-assisted mapping and data transformation. AI can analyze multiple sources of information -such as XML files, Excel mapping sheets, or Word specifications – and propose a unified data structure.
Once the proposal is revised it can be adjusted – by for example adding fields:
“Add company code in Idoc” will result in adding BUKRS field in the structure – AI understands the context and uses SAP technical name for it.
Something very similar happens when we try to add company code field to Salesforce API object – it understands that we are working with custom fields and suffix “__c” will be needed.
Once source and target messages are confirmed – the next step is to generate mapping definition. Users can upload the definition in whatever format they have (like Excel, PDF, CSV) and the workflow will use them to propose how messages can be mapped in a visual way. After it the technical object (like in this case groovy script) for SAP Integration Suite can be generated. This highlighted one of the most practical and immediately valuable applications of AI in SAP Integration: saving significant time on complex mappings while keeping developers in control. This highlighted one of the most practical and immediately valuable applications of AI in SAP Integration: saving significant time on complex mappings while keeping developers in control.
Tenant Interaction and Management
The webinar also explored AI as a conversational interface for managing SAP Cloud Integration tenants.
Capabilities demonstrated included:
- Listing all integration packages.
- Retrieving monitoring data, such as the top iFlows with errors.
- Explaining what a specific iFlow does.
- Performing where-used searches across the tenant.
A future possibility discussed was integrating AI with an organization’s knowledge base, allowing it to leverage existing documentation or even update it automatically after deployments.
Security and Architecture
Security was a recurring theme throughout the session. The main concern was how to make sure sensitive integration data doesn’t get exposed to public AI models.
Proposed solutions included:
- Hosting AI models within SAP BTP AI Core, ensuring all processing stays inside the company ecosystem.
- Using specialized, small language models (SLMs) trained for specific tasks such as BPMN generation.
The demonstrated prototype uses a secure architecture, where the LLM never directly accesses the tenant.
All tenant queries go through a Python backend that performs controlled API calls and enforces role-based authorization before any request is made.
From experiments to product
During the session, Radek shared how his experiments have evolved into something more practical. AI assistant for SAP Integration Suite – isVisible – that is now being developed by Sygeon team – that helps with repetitive tasks, provides insights on integration objects and runtime, and makes everyday development work faster and more efficient.
The feedback from the webinar was very positive, so a preview version of the application is being prepared.
You can already sign up for early access here:
The Future Role of the Integration Developer
The webinar offered a grounded, practical view of how AI is reshaping SAP Integration.
- AI can accelerate development and analysis through automation of repetitive tasks.
- The best results come from a human-guided, multi-step approach – not single-prompt generation.
- Security and governance must be built in from the start.
- Junior-level tasks, such as documentation or basic transformations, may become automated.
- Business users can leverage AI for simple automation, but complex, enterprise-grade integration will remain the domain of skilled developers.
If you missed the session, you can also watch the full recording here:
