To date the only AI that I've seen deliver any sort of semi-useful work in the corporate world has been meeting summarization technology. Basically the AI attempts to interpret what was said in the meeting in order to deliver a summary.
I call it semi-useful because it doesn't understand nuance, varying slang terms versus official terms for industry-speak, and it can't even handle wisecracking.
I suppose that in a forensic-ish role it could help because it could analyze large datasets to find patterns, datasets that are so large that it's difficult for humans to evaluate all of the conditions, but whether or not this actually happens in a corporate role is hard to say. Most of the really huge data-processing AI systems are being overtly managed by developers. The biggest companies might be able to self-manage this, and other large companies not in this line of work might well sub-out this systems maintenance to technical services companies to maintain it on their behalf, but whether or not this trickles down to smaller companies would very much depend on how much it costs, and really how good the results actually are.
If I was middle-management I would be very cautious about embracing AI. The -GPT systems of the world have already demonstrated how utter crap they are sometimes, and my guess is that the sort of AI that will be available to them to potentially replace team members will be more like that and less like big-data AI. Those middle-managers will find their own roles diminished if AI comes in like this CEO thinks it will, and that not only threatens these middle-managers' positions through garbage-out, it also threatens to turn these middle-managers into the frontline white-collar workers again as they have fewer and fewer people to supervise and are now just keyboard-monkeys themselves.
For Ford and other manufacturing companies, I expect they will continue to push for savings in the manufacturing side of the house more than the administration, marketing, sales, and management sides of the house. I expect that they'll use it in the combination of design and manufacturing to attempt to produce product designs that require fewer and fewer people to be involved to actually manufacture said products. This is particularly an issue for automakers where their contracts for manufacturing labor might require them to pay workers when the plants are idle because the plants are being retooled for different design or because the company mispredicted sales forecasts and overbuilt and needs to idle until inventory is reduced. I could see them wanting to reduce the number of actual workers because then they don't have to contend with labor considerations for manufacturing tasks that don't involve humans. But that may not even be a matter of AI, that may just be more white-collar engineers working on how to design for the factory even more than they do today.
How much of this sort of announcement by corporate leadership could be attributed to misdirection? Threaten the positions of the office workers closest to them to distract while laying off the manufacturing workers at the far-flung plants?