Is Artificial Intelligence Determining Future Purchasing?
This article is the long version of the third part of a six-part series currently appearing in DVZ (Deutsche Verkehrs Zeitung, dvz.de) that analyzes the future of each individual node in the supply chain. Together with SAP, Prof. Hoberg and his team spent the last year investigating how supply chains are likely to develop in the future, drawing on contributions from 660 supply chain experts.
The third part of the joint study by SAP Business Consulting and Kühne Logistics University on the supply chain of the future deals with purchasing. We do not assume the future of purchasing looks the same for all companies. Rather, key trends and key technologies play a big role. Their impact varies greatly depending on the industry and business model.
The shaping trends for purchasing (as well as for the other parts of the supply chain) include the global economic situation and geopolitical tensions, the growing importance of sustainability, the scarcity of raw materials, and the changes for the people involved. These issues are not new, but they will play a major role in the coming years and will sometimes become even bigger.
Key technologies are in the focus, e.g. integrated transactional purchasing applications with new semantic data models, AI agents, AI-based data analyses, and insights from the evaluation of information from business networks. Artificial intelligence is not the only driving technology but will become increasingly relevant in the coming years. No process in modern purchasing organizations will remain unaffected by its influence. Against this background, purchasing organizations have to work out which trends and technologies are particularly important for their company to plan the path to a successful future.
The study conducted an online survey of 174 supply chain managers with people responsibility (April 2025). There was a clear discrepancy between ambitious visions for the future and the sad reality. On the one hand, 28% of respondents expected purchasing managers to play an important role in shaping corporate strategy and security of supply in 2035, while 31% expected buying organizations to achieve significantly better AI-driven outcomes with high quality data. On the other hand, only 12% say that procurement in their organization is now fully integrated into C-level decision-making, and only 3% think AI-powered purchasing analytics are great today.
How should this discrepancy be dealt with? And how to deal with the uncertainty of every look into the future? The clear recommendation for all purchasing organizations is to do their homework first, to take decisions and to perform a couple of no-regret moves. It is clear that only purchasing organizations following these five important points are optimally positioned for the coming years.
Firstly, clearly defined strategies and responsible people are in place for the most important categories and sub-categories. These category strategies must be kept up to date in suitable applications instead of gathering dust in presentations stored on shared drives. The strategies will be the trigger for activities, and the focal point for monitoring the progress of the activities.
Secondly, the operating model of the purchasing organization must be optimized by shifting personnel capacities from administrative activities to value-adding tasks. The goal is to work with a small but highly skilled workforce.
Third, clearly defined and standardized processes are required that exist in Business Process Model and Notation (BPMN) notation and are mapped in a process management solution, not a drawing tool. On this basis, simulations and comparisons of as-is and to-be can be carried out. Also, the processes can be linked to business requirements, test cases, application documentation, driving AI based automation in these related activities.
Fourth, modern purchasing applications should be used to support end-to-end processes, without media breaks and with a steadily growing degree of automation. An end-to-end flow of data and information is essential. Specifications and prices belong into the apps themselves, not in file attachments. These pieces of information are critical parts of the data that drives AI. Similarly, the apps must be integrated so data is not transferred manually or via bots.
Fifth, all relevant data from various procurement apps and third-party data must be available for analysis. The key is that this data is available on an appropriate platform and mapped to a semantic model. This is the only way to get an overview of the entire purchasing process, to derive in-depth insights and to drive more AI use cases.
The last point is of particular importance. Only if data on purchasing processes is available in structured form, across all purchasing processes, in good quality, will the purchasing organization have a solid basis for using AI functionalities of any kind, especially AI-driven analytics and AI agents, which can take on complex tasks on their own.
AI agents sound a bit like science fiction for many persons in the procurement space – that is why the figure below outlines an agentic AI scenario in strategic procurement. The CPO sets a task for a procurement orchestration agent which in turn sets tasks for other, more specialized procurement agents (e.g. spend analysis, sourcing activities, contract creation) to achieve an overall goal, the optimization of a specific sub-category.
Similarly, all other parties involved in the purchasing processes can be supported. Requestors can interact with an operational purchasing agents to quickly and clearly describe their needs. That operational purchasing agent reaches out to other agents leveraging existing contracts or catalogs, adhering to all purchasing guidelines, and covering topics such as account assignment, tax determination, and transportation costs. With the help of their own agents, suppliers will find it easy to participate in complex tenders, bid in Dutch or Japanese auctions, negotiate and sign contracts, and provide price lists and catalogs in specific formats to the purchasing organization. Since suppliers will use their own AI agents, there will be an increasing frequency of interaction between multiple agents.
In summary, it can be said that those purchasing organizations will be optimally positioned for the future, who correctly classify the relevance of future trends and technologies to their own business, and who conclude the five no-regret moves mentioned above.
This article was written in collaboration with Prof. Kai Hoberg (Kühne Logistics University) in the context of the SAP-KLU study "The Supply Chain of the Future”.
