The AI Summit which took place in Paris earlier this month, swiftly and powerfully shifted the tone around AI toward big investment, calculated risk taking and the drive to win with the best AI development and real-life outcomes for business and society. It should provide another catalyst for manufacturing leaders to think more about their own AI investment strategy, even in the face of tough times, fluctuating demand and geopolitics. Rudolf Schambeck, Manager, Machine Vision, Zebra Technologies, explains more.
The summit’s EU AI Champions initiative will be supported by a range of companies and pledges of €150bn investment plus €50bn from the EU for AI in Europe. Meanwhile, France is promised €109bn investment for its AI ecosystem, and the UK announced a new AI growth plan, £14bn and over 13,000 jobs by tech firms on top of a £25bn investment previously announced.
Manufacturers’ data conundrum
The EU President, Ursula von der Leyen said: “European AI focuses on AI adoption in complex applications, using our unique industrial and manufacturing data and know-how,” which is true. The use cases and compliance requirements for pharmaceuticals, automotives and food are high and require more advanced AI such as deep learning and 3D scan software to handle them.
These AI ambitions are naturally tied to the need for data – lots of high value, good quality data, which we see in manufacturing across many processes. “Industries will be able to collaborate and federate their data. We are creating the safe space for them… because AI needs competition but also collaboration,” said the von der Leyen. And a letter by founders and CEOs called for high-value data sets that could be publicly accessible in privacy-preserving and safe ways, and the UK Prime Minister has proposed a national data library that researchers, non-profits and others can access.
This reminds us of the challenge manufacturers face within their own companies. A recent piece of research by Zebra Technologies found that among machine vision leaders in the automotive industry, almost 20% in Germany and the UK say their AI machine vision could be working better or doing more. For AI solutions to achieve their potential, the question of data must be addressed.
The volume of data being created at the edge of business can be turned into value. That could be data for training and testing AI models or acting as feedback to refine processes around manufacturing and inspection. Once data and AI are integrated, the path to process automation – with smart cameras, sensors, and vision-guided robotics – becomes a reality, allowing leaders to reallocate valuable frontline workers to growth-focused areas.
However, manufacturing sites and regions can operate in silos, with little to no sharing of data, even for identical or similar workflows. Experience and time available can vary between teams and sites which can make achieving data quality more challenging, compounded by struggling to hire the right talent with the right skills and experience.
Data needs to be stored, annotated and used for training models in a consistent way, with other data sets needed for model testing. It makes no sense for company data to remain siloed, to the detriment of better training for AI.
But how are manufacturers to achieve this if they can’t leverage all the data available to them across sites, countries and regions? How much growth potential and workforce productivity are being lost? The hesitancy around the cloud due to privacy, security and intellectual property needs to be overcome. A cloud-based solution would allow users to securely upload, label and annotate data from multiple manufacturing locations across site, country and region and deliver scalability and accessibility of computing power.
The options for getting AI solutions working better and doing more are available, whatever the use case might be. There is software, cameras and sensors for use cases including electric battery and semiconductor inspection, fresh food sortation, packaging compliance and quality, serial number and character reading, and defect detection for automotive parts and finished items.
However, AI working better or doing more needs to be measured with appropriate timelines, return on investment metrics and quality data. Alongside data management transformation, intelligent automation and greater asset and inventory visibility require appropriate implementations, workforce training and operational adjustments which take time. Other AI solutions take a low/no code approach, and come ready out of the box, delivering a faster return on investment.
In praise of middle management
Who can shift the tone and lead the drive to win when it comes to AI within manufacturing companies? Recent research by McKinsey highlighted that many millennials aged 35 to 44 are managers and team leaders in their companies i.e. middle managers. They are on the frontline of business, walking the factory floor and acting as the bridge between the frontline workforce and senior leadership. These middle managers report having the most experience and enthusiasm about AI, with 62% of employees between the ages of 35 and 44 reporting high levels of expertise with AI.
Separate research found that only 30% of business leaders (CEOs, presidents, other C-suite executives, senior vice presidents and executive vice presidents) increase resourcing for growth initiatives in core, adjacent or new businesses during periods of volatility. Only 29% said they invest 30% or more of their time on long-term growth initiatives.
Middle managers are a key talent pool senior leaders should leverage to answer questions like how AI, such as deep learning machine vision, can help accelerate through a challenging market and drive long-term growth, elevate productivity, automate processes and improve quality. Despite the initial reaction to contract and pause projects in the face of hard times, now might be the right moment to rethink the role middle management plays when it comes to prioritising growth and acting boldly with AI and machine vision.
For more articles like this, visit our Industrial Data & AI channel.
