My name is Jonathan Wieder, I am a dual student at SAP SE studying Business Informatics at DHBW Mannheim. During my practical phase in the product management for Enterprise Asset Management – supervised by Raik Kulinna, I worked on a project that aimed to optimize internal product management processes at SAP through the integration of generative AI — and I’m excited to share the outcome with you.
In this blog post, I will take you through the development of a macOS-based prototype that enables product managers to automate text-based tasks — such as rewriting, translating, summarizing, and generating user stories — using Apple Shortcuts and SAP Business AI. The goal was to embed generative AI seamlessly into the daily workflow of product managers while ensuring SAP’s high compliance standards for AI usage.
At SAP, product managers work in a dynamic international environment, balancing tasks such as defining user stories, developing personas, and tailoring communication for diverse stakeholders. While generative AI holds significant promise for accelerating these activities, strict internal data and compliance policies prohibit the use of external AI tools.
To address this challenge, SAP AI Launchpad turned out to be the ideal solution. It offers secure, OAuth2-authenticated access to internal large language models (LLMs), including OpenAI’s GPT-4o, ensuring compliance with internal data protection standards. The solution had to be fast, seamlessly integrable into macOS workflows, and fully compliant with SAP’s internal security requirements—making SAP AI Launchpad the best fit for the job.
Component | Technology |
Interface | Apple Shortcuts (macOS context menu) |
Authentication | OAuth2 Client Credentials Flow |
AI Engine | SAP AI Launchpad with e.g. GPT-4o |
Script Layer | JavaScript for Automation (Apple JXA) |
Tested Use Cases | Rewrite, Translate, Summarize, Persona & User Story Generation |
Each functionality is triggered via macOS context menu on selected text:
Each request is authenticated via OAuth2 and sent to the SAP AI Launchpad via HTTP POST, with responses returned and optionally inserted back into the original app context.
var aiRequestBody = {
"model": "gpt-4o",
"messages": [{ "role": "user", "content": prompt }]
};
var aiCommand = `
curl -s -X POST \\
-H "Content-Type: application/json" \\
-H "Authorization: Bearer ${accessToken}" \\
-d '${JSON.stringify(aiRequestBody)}' \\
'${aiUrl}'
`;
var aiResponse = app.doShellScript(aiCommand);
The AI shortcut can be accessed via the context menu. This menu is By right-clicking on a text, the context menu opens, where all shortcuts within the service submenu are displayed. Apple provides a native way to handle menu input and pass text between steps.
Example Workflow:
The distribution of shortcuts is also simple: via sharing of a link or through an exported file.
The prototype was validated against seven user stories defined via a Value Proposition Canvas. All functional requirements were fully met. The solution:
This project demonstrates how product managers at SAP can benefit from embedded generative AI without leaving their tools or violating AI policies. Using Apple Shortcuts and SAP AI Launchpad, a lightweight yet powerful assistant can support everyday work from writing user stories to translating documentation.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
47 | |
24 | |
18 | |
11 | |
10 | |
7 | |
6 | |
5 | |
5 | |
5 |