Red Sofa: BMW's Oliver Ganser on transforming supply chains with agentic AI

Описание: As well as delivering a keynote session on scaling next-generation AI and taking part in a panel discussion on optimising supply chins with genAI and agentic AI, Oliver Ganser, vice president of digitalisation of the purchasing and supplier network at BMW Group, also sat on Automotive Logistics' Red Sofa at ALSC Digital Strategies Europe to discuss these topics further.

Human-AI cooperation
Ganser described the emergence of genAI as "a matter of uncertainty", particularly surrounding how the technology will impact the day-to-day lives of those working in the automotive industry and its wider supply chain.

"We as leaders have to give certainty that [people] are the core asset of our operations," he noted, adding that humans will not be replaced by such technology.

He explained that it's important to clearly outline what agentic AI can allow people to accomplish, and clearly outline what can be delegated to an agent and what should not be.

Agentic AI in practise
Both on stage and on the Red Sofa, Ganser shared a fascinating use case of agentic AI that has allowed BMW to revolutionise its inventory and purchasing processes. What was previously done by 850 purchasers and required constant attention and human-to-human communication, now is led by an agent and a human is only looped in when there is a deviation from the norm or a critical escalation is needed.

"So far, 80% [of the processes] are run completely by the agent and only 20% of the case needed a human in loop for specific cases," he shared. Cutting down the number of "cumbersome" processes for which humans are required is something Ganser believes would be welcomed by the workforce, but he noted that it is up to those in leadership positions to effectively communicate the benefits of this technology across the company.

Improving communication throughout the supply chain with AI
Ganser also shared that agentic AI has the potential to streamline communication not just internally, but also externally – improving the efficiency of communication with external suppliers and supply chain partners. "I strongly believe that there is a lot of communication between two or more companies in a value chain that is completely unnecessary and doesn't get any value," he remarked.

Improving communication throughout the supply chain with AI
Ganser also shared that agentic AI has the potential to streamline communication not just internally, but also externally – improving the efficiency of communication with external suppliers and supply chain partners. "I strongly believe that there is a lot of communication between two or more companies in a value chain that is completely unnecessary and doesn't get any value," he remarked.

Rather than spend time writing emails and filling out forms, Ganser shared his belief that an agent could guide people through these processes. But to get to this stage, he identified three key attributes that a supply chain must have: digital trust; access and control rights; and semantic ontology in data.

"I strongly believe that inter-company connections and business optimisation in the future, if we want to do it right, is done by agents," he added.

Ganser, as head of industry consortium Catena-X – a collaborative data ecosystem for the automotive industry, also emphasised the need for a "common data foundation" to establish the correct environment for Agentic AI to deliver the greatest results for supply chains.

"I think Catena-X is the only way forward as an industry to lay a foundation that then with serve as an AI foundation for agents," he said. "We can as a second step build on top of that, but without that we will fail."