Ford Motor Co. is integrating artificial intelligence (AI) across functions--be it in the design studio or on the factory floor--to turn “manual hours” into “pushes of a button,” while keeping human expertise firmly in the driver's seat, the 120-year-old automaker's chief data, AI and analytics officer Franziska Bell told ET. The use cases span design, engineering, quality assurance, customer support, data management, and software development, and most of these cutting-edge advancements are being through teams in Chennai and Bangalore.
“Ford has already deployed artificial intelligence at scale across our entire product life cycle,” Bell said. “Our strategy is really human AI teams. We believe in augmenting humans through artificial intelligence. The human is very much still in the driver's seat and can move forward at a much faster pace and also provide more variations than would have been possible previously.”
She said that artificial intelligence is not yet advanced enough for full automation of entire jobs and still faces challenges such as hallucinations.
“I believe in the ingenuity of humans,” Bell said. “The strategy we have developed therefore is really a collaboration where we want to harness the best of both worlds--human ingenuity, subject matter expertise (and) logical reasoning, with the computational power of artificial intelligence, which is very good at finding patterns in very large volumes of data.”
The Dearborn, Michigan-headquartered company has developed use cases that include co-designing a car, going from a manual sketch to a 2D or 3D rendering in the blink of an eye.
“We can generate hundreds of variations at a push of a button and this really elevates our designers to executive directors,” Bell said. “We do this not only for the exterior but also for the interior (of the car). We do this also for intricate wheel designs as well. Traditionally, you would start off with a small human-drawn sketch.”
That drawing may not be required either.
“We even built a tool that makes this sketch obsolete,” she said. “If people want to start off with a natural language query, there is a tool now that people can just type in (their specifications) and then literally click a button, generate, and then you get various different vehicle designs as well.”
AI is also used in engineering and testing, with proprietary models developed for aerodynamic drag computations.
“This is one of the key bottlenecks in the auto manufacturing design and development process,” Bell said. “Previously, the simulations took about 16 to 18 hours to complete. And now with this deep learning approach, we can do this in 10 seconds. That's about a 5,000-fold speed up. Again, very much with the human expert in the loop, but really helping facilitate a much faster iterative cycle.”
In areas such as quality assurance, the company uses computer vision. Last year alone, Ford made 160 million inspections using this, implementing it in over 30 plants globally, she said.
“This really helps us to identify any potential quality issues before the car rolls off the plant including very small millimetre deviations,” Bell said. “We also have built a new unified experience for the customer so instead of having information siloed in different areas like user manuals, warranty documents and service documents, this is now all accessible in a single chat interface.”
The company has also developed a supply chain risk assist tool to identify problems early enough and work with partners to mitigate them.
“It's obviously a big opportunity for Ford and a phenomenal business problem to work on,” Bell said. “But secondly we're also using some cutting-edge technology here--a multi-agentic system, which has been on top of people's minds for some time, but I think we are very much at the forefront at launching these capabilities at scale.”
Part of the team located at the Ford Business Solutions office in Chennai works on this solution. The site is Ford’s biggest global capability centre (GCC) worldwide.
“Ford has already deployed artificial intelligence at scale across our entire product life cycle,” Bell said. “Our strategy is really human AI teams. We believe in augmenting humans through artificial intelligence. The human is very much still in the driver's seat and can move forward at a much faster pace and also provide more variations than would have been possible previously.”
She said that artificial intelligence is not yet advanced enough for full automation of entire jobs and still faces challenges such as hallucinations.
“I believe in the ingenuity of humans,” Bell said. “The strategy we have developed therefore is really a collaboration where we want to harness the best of both worlds--human ingenuity, subject matter expertise (and) logical reasoning, with the computational power of artificial intelligence, which is very good at finding patterns in very large volumes of data.”
The Dearborn, Michigan-headquartered company has developed use cases that include co-designing a car, going from a manual sketch to a 2D or 3D rendering in the blink of an eye.
“We can generate hundreds of variations at a push of a button and this really elevates our designers to executive directors,” Bell said. “We do this not only for the exterior but also for the interior (of the car). We do this also for intricate wheel designs as well. Traditionally, you would start off with a small human-drawn sketch.”
That drawing may not be required either.
“We even built a tool that makes this sketch obsolete,” she said. “If people want to start off with a natural language query, there is a tool now that people can just type in (their specifications) and then literally click a button, generate, and then you get various different vehicle designs as well.”
AI is also used in engineering and testing, with proprietary models developed for aerodynamic drag computations.
“This is one of the key bottlenecks in the auto manufacturing design and development process,” Bell said. “Previously, the simulations took about 16 to 18 hours to complete. And now with this deep learning approach, we can do this in 10 seconds. That's about a 5,000-fold speed up. Again, very much with the human expert in the loop, but really helping facilitate a much faster iterative cycle.”
In areas such as quality assurance, the company uses computer vision. Last year alone, Ford made 160 million inspections using this, implementing it in over 30 plants globally, she said.
“This really helps us to identify any potential quality issues before the car rolls off the plant including very small millimetre deviations,” Bell said. “We also have built a new unified experience for the customer so instead of having information siloed in different areas like user manuals, warranty documents and service documents, this is now all accessible in a single chat interface.”
The company has also developed a supply chain risk assist tool to identify problems early enough and work with partners to mitigate them.
“It's obviously a big opportunity for Ford and a phenomenal business problem to work on,” Bell said. “But secondly we're also using some cutting-edge technology here--a multi-agentic system, which has been on top of people's minds for some time, but I think we are very much at the forefront at launching these capabilities at scale.”
Part of the team located at the Ford Business Solutions office in Chennai works on this solution. The site is Ford’s biggest global capability centre (GCC) worldwide.