When it comes to AI, there are so many possibilities that it can be difficult to know where to start. And that’s why real-world examples of AI in production are so valuable. Here’s a cool one from the SAP Innovation Awards 2026: Promocean
Imagine that you are a world-class brand such as Heineken, Lindt chocolates, Bacardi or Lacoste and you are organizing a promotional event. You want the event to feel special, different, creative, but still 100% aligned with your brand image and of course, for the right price and delivered on time. That’s when you might turn to Promocean, the winner of the Technology Pathfinder category of the SAP Innovation Awards 2026.
The Dutch company is trusted by some of the world’s biggest brands to create award-winning packaging and point-of-sale marketing materials.
Procurement is crucial for Promocean because they need to efficiently source highly customized materials and services from around the world, while meeting strict cost, quality and delivery requirements.
Surveys show that communication between suppliers and buyers is a major source of frustration in almost every industry. Promocean had an overloaded team spending up to 19 hours a day tracking purchase orders, using manual email and Excel-based processes that simply didn’t scale.
Additionally, delays in the supply chain were not as visible as they should be and there was no easy way to track the reliability of different suppliers. and when problems arose, it was difficult to collaborate between different internal teams.
At an AI enterprise hackathon organized late last year by SAP Netherlands and Dutch SAP user group VNSG, Promocean saw an opportunity to use AI to reduce these risks and inefficiencies. Working with their long-term strategic consulting partner SOA PEOPLE and an expert SAP coach, they designed and implemented an AI-powered procurement solution powered by SAP Business AI Platform.
The way it works is interesting. First, there is the design phase using AI. Promocean had a big advantage: they had already thoroughly documented what the process should look like. Then, using the SAP AI platform, they fed it into a GPT model using a now-standard RAG approach, augmented generation retrieval, and used it to create a very sophisticated and complex rules-based engine to handle different types of queries.
Then there is the operational phase: an unattended bot using SAP Build Process Automation receives Outlook emails every 30 minutes. They are then interpreted using large language models from Mistral and Anthropic in SAP’s GenAI Hub and enriched with back-end data from an on-premises SAP ECC system, for example by looking up the current status of an order or invoice.
Demonstration of the solution in action.
The email summary is then passed to the rules engine to determine the best next step and generate intelligent action proposals. Users can then review and manage these proposals in real time, using the latest information available, using a custom Purchase Order Control Center built with Fiori and SAP’s cloud application programming model in the SAP Build Work Zone.
The final choice always lies with the end user. If they accept the action, the system can perform the next steps automatically. If they do not agree with the proposed solution, that feedback is returned to the system to improve it over time.
The solution went live on January 19, 2026 after less than 20 weeks of development, resulting in more automated, timely and consistent tracking of purchase orders, reducing manual effort by 63% and with fewer errors. The team estimates that by leveraging AI techniques they were able to reduce development effort by 90%. There is faster and more consistent tracking of suppliers as the team can move from repetitive operational work to focusing on what is important. And the system is scalable and future-proof, a platform for other AI projects, providing value today while also being ready for the move from Promocean to S/4 HANA at some point in the future.
Lessons learned:
- Look for opportunities to participate in hackathons and design workshops with your suppliers and strategic partners to find the best. most feasible AI projects for your organization and be sure to engage the employees involved to leverage their knowledge and enthusiasm.
- Real-world AI works best when it’s built on a solid foundation, like the SAP Business AI platform, which helps you move from prototype to production and prepare for the future of AI.
You can get more details about the solution in Promocean’s SAP Innovation Award entry and in this video by Twan van den Broek: