Companies are spending billions of dollars on artificial intelligence to conduct tasks that range from clerical to mechanical to planning to operating humanoid robots on factory floors.
Indeed, AI seems to be everywhere, but in many cases its abilities aren’t showing up on the bottom line in black ink.
Almost a third of the 4,454 CEOs surveyed by financial services company PwC reported increased revenue from AI in the past year and a 26% said they are seeing lower costs, however more than half—56%–said they’ve realized neither revenue nor cost benefits from $40 billion in enterprise investment in AI.
While the CEOs involved in that survey released in January encompassed a wide breadth of industries, the results are indicative of the challenge facing automotive manufacturers and suppliers, according to assessments from a number of insiders and experts.
“Automotive is really a microcosm of what we’re seeing across all industry,” declared Ajay Chawla, CEO at OnTrac AI, a Bloomfield Hills, Michigan AI and automation implementation and consulting firm.
Both automakers and suppliers are conducting pilot programs using AI but often thos programs are fragmented and siloed across many different divisions in the company, Chawla explained, in an interview.
An example he gives is if customer service is piloting one AI tool while human resources and sales and marketing are piloting others creating a fragmented approach—a scenario that’s showing up more often in the auto industry.
He suggests a more “inclusive” roadmap across an entire company led by the C suite, much the way Ford Motor Co. CEO Jim Farley has said the automaker is going to invest in AI but with direction from the top executive levels down through the organization.
“A lot of these companies that aren’t getting it right are lower level people that are trying to put AI in themselves, working on pilots and they’re picking tools that don’t jive with the rest of the company,” asserted Chawla. “There are a thousand different AI tools you can go out and find that are off the shelf, and you can go buy, and every every one looks great to the naked eye, but if you can’t implement them and have a streamlined AI strategy across the corporation, you’re not going to come out ahead.”
Executives and engineers at several automotive suppliers wrestled with this very challenge during a panel discussion sponsored by the Society of Automotive Analysts at the suburban Detroit headquarters of engineering firm Munro and Associates last week.
When asked to respond to the PwC survey findings, there was a combination of surprise, skepticism and, in one instance, just a bit of defensiveness.
“We all have to learn the payback with the availability of the tools. On top of that, they’re rapidly evolving,” said Armin von Czarnowski, technical account manager, Munro & Associates. “So, what we’ve all seen is powerful six months ago, three months ago changes. So, you’ve got to give a little bit of time.”
Indeed, even though AI seems to have permeated even the most mundane functions, a top executive at engineering firm Pratt Miller suggested familiarity with the technology remains a challenge to some.
“I think there’s initially a fear of AI, to be honest, where I hope we’re past that with most people–I don’t even know what it is, or how to leverage it,” said Pratt Miller vice president of mobility Jon Honig. “But the biggest thing I think companies are struggling with is, how do I incorporate this into my business model?”
Despite the enthusiasm and anxiousness for some companies to incorporate AI, Scott Walter, robotics research diligence director, at RoboStrategy, believes payoff will come with patience, along with adequate due diligence.
“There’s one thing when we see in the press and the hype that’s out there, but when you actually talk to the engineers and people on the job they have more realistic sense of what the timelines are, but they realize it will be productive,” said Walter. “But if you try to do something way too complex, way too early, you’re going to end up with the same result.”
At robotics company FANUC America, success came after repeated testing and analysis, according to Amar Dhaliwal, general manager of the company’s automotive systems group, who revealed, “I’m seeing it real life happen. We started around 70% success. We got to 99.3 just by running models and models, and then we put it in a physical world.”
One key to companies in the auto industry and beyond realizing payback for their AI investment is making sure human beings remain part of the process learning how to work side-by-side with the technology, according to Chawla, declaring, “to just give AI something to do and say set it and forget it, and we’re not going to worry about it, is absolutely crazy to do now.”
Amid doubts, skepticism and impatience regarding AI’s payoff, sometimes it just takes some words from the academic world to calm the jitters and level-set expectations.
In this case, it’s the oral balm from Andrew Watchorn, a lecturer Kettering University, which offers degrees in automotive engineering, who advised the panel’s audience, “even though you have not seen the return of investment yet, yes, it may be too soon. We may still be at the top of a hype curve, and there may be a valley later on, but people are starting to see the actual quantifiable difference between someone who is equipped and has good judgment using this.”
