How to bring Industrial AI to real tasks
Artificial intelligence has a lot of potential when it comes to changing the way people work yet all to frequently, realizing these gains means radical adjustments to existing processes, retraining or a general upending of the status quo. While there is value in change, equally there is value in streamlining adoption as well. For AI, aiming for seamless integration with existing processes will be an important first step in long-term adoption.
In a recent podcast, host Conor Pieck is joined by guests Matthias Loskyll, Senior Director of AI and Robotics at Siemens Digital Industries and Samir Desai, Senior Director and Global Program Head for Data and AI at Siemens Digital Industries to explore the ways AI is being integrated with existing processes from design to manufacturing and beyond.
Check out the full podcast here or keep reading for a summary of the highlights from that conversation.
Engineering AI brings AI to the user
A key element to successful AI adoption, especially in a complex environment, is ease of use. Samir highlights how, when bringing AI to many of their tools, the intention is to bring the AI models into existing tools/apps in concept he calls Engineering AI. Rather than using large, general purpose AI models that may not integrate well with existing tools and workflows, Engineering AI uses purpose-built AI models integrated directly with the tool itself to offer a more seamless experience.
Getting the most out of Industrial AI really means applying the right tool to the right job. While general purpose LLMs may work well enough across a broad range of tasks, they rarely excel in any particular area. When it comes to industrial tasks, which often have little training data and stringent requirements, these general purpose models can quickly fall behind purpose built ones, not only offering lesser benefits, but being more difficult for users to learn and operate.
The impact of AI on the shop floor
Whether bringing AI to the design process or the shop floor, there are many challenges that must be overcome. While some are unique, in many areas they are the same. User adoption and buy-in being one key area where AI must prove its worth to be taken seriously. Matthias explains how, in bringing AI to the factory, the goal was, much like for Samir and his team, to create AI that was easy to use and seamless for the end user.
In factory operations, reliability is king which can, unfortunately, result in the slow adoption of new technologies. To address this, bringing AI to the factory means creating tools and models that are not only robust and reliable but offer clear improvements over existing methods all while being easy to use and seamless to integrate. While this may seem like a tall order, Matthias says Industrial AI solutions for quality inspection and predictive maintenance are already in the hands of customers, offering a seamlessly improved experience while retaining the required level of robustness. Simple and effective AI tools like these will serve as the building blocks for AI in factory operation going forward, laying a foundation of trust on which more complex applications can be built.
The importance of data fabrics
Purpose building AI for industry makes a lot of sense yet, to achieve that goal requires vast amounts of data. Compared to general purpose models which have access to vast amounts of publicly available data, creating AI models for industry specific tasks requires data that is far more difficult to acquire. This is where data fabrics come in.
Data fabrics allow siloed industrial data to be connected into a single data lake but, crucially, embed context and metadata that is important for training AI models to understand industrial tasks and data. Deploying a data fabric is as much a part of the process of using Industrial AI as building the tool itself in order to ensure the models are capable of handling industry specific tasks appropriately while having access to the freshest possible data.
AI has a lot to overcome to reach widespread adoption in industry but the biggest element is the human factor. Building AI in such a way that it can offer seamless support to humans users, without requiring long and difficult retraining, to get real value, will be the foundation on which true Industrial AI can be built.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
