Observations from the ABB Innovation Connect Day, ABB Campus Mannheim, October 1, 2026
On October 1, ABB hosted eleven start-ups for a pitch session at its Mannheim campus, attended by around 25 employees from various divisions of the company, in collaboration with the accelerator Up2B. This was followed by speed-dating sessions to determine which conversations might lead to concrete collaborations. oee.ai was one of the eleven companies involved.
The theme of the afternoon can be summarized in one sentence: Artificial intelligence has arrived in our products.
What ABB was looking for
ABB had categorized the participants into three areas in advance: energy efficiency, process efficiency, and seamless integration. This classification says a lot about where a corporation of this size seeks value. Underlying these topics was the question of what ultimately reaches ABB's customers. The pitches were measured against this standard. At ABB, innovation is described as part of the company's core identity rather than a special project separate from day-to-day business, and this mindset defined the afternoon.
Where the solutions draw their information from
More interesting than the categorization was what the presented solutions were actually built upon. The majority read documents, specifications, master data, or source code to extract requirements from specifications, consolidate part master data in procurement, generate test cases for embedded systems, or identify vulnerabilities in repositories.
The decision-maker is often found in development, procurement, or IT—areas where the benefits can be demonstrated quickly.
AI is no longer a differentiator
On this afternoon, artificial intelligence was truly embedded in almost every product, predominantly in the form of language models and agents. It was the exception for a provider to merely announce AI. This shifts the question that an industrial company must ask a provider. Today, it is: what data is the AI calculating on, how reliable is this data, and who is responsible for the result?
What we pitched
Our contribution focused on a loss at a machining center, tracked from the moment of downtime to the documented countermeasure: a tool breaks, and the machine stops. The operator describes what happened in their own words at the terminal. The AI then suggests countermeasures that have been previously learned and approved. This reduces the mean time to repair and, consequently, increases OEE. Additionally, knowledge is shared across shifts and is preserved even when employees leave the company.
We proposed this use case to ABB as a pilot project: two machines, eight weeks, from installation and data collection to the AI-generated suggestion.
Where it all comes together
Anyone wanting to use AI in manufacturing must first solve the data question. Availability, downtime, and losses occur at the equipment level, and the information about them lies within the controllers and with the people standing at the machine.
During one of the discussions, a comment was made that hits the nail on the head: the drive is the first principle of OEE. Equipment productivity begins with the question of whether something is moving, and a software-driven drive knows this state the moment it occurs. ABB builds exactly these drives and the associated controllers. Therefore, an ABB product is at the beginning of many productivity calculations.
However, the drive does not know why the machine is stopped. The reason is often known by the controller—also an ABB product—and where it doesn't know, the person at the machine provides the answer. oee.ai turns this into a structured root cause with a countermeasure that someone can look up the next time an incident occurs. The two ends of this chain belong together, and that is exactly what we can build on together. The speed-dating sessions after the pitches provided the starting point for this.
Thanks to Up2B for the organization and to ABB for an open and productive afternoon.