If AI spending is only improving the bottom line via job cuts, you’re doing AI spending wrong. You’re misinterpreting gains from reduced payroll expenses as increased productivity or market share. What a fucking mistake. If AI spending isn’t directly contributing to sales in some fashion, then you’re just clowning around by pushing money from payrolls to revenue. Your company hasn’t innovated, your product hasn’t become more attractive or competitive, and your business is just as profitable as it was prior to the AI spending.
Think about a startup that previously would have take 3 or 5 people to launch, and was not viable. But now can be launched with 1 person and be viable. big business also usually has similar internal startups or new ventures with similar economics.
> If AI spending isn’t directly contributing to sales in some fashion, then you’re just clowning around by pushing money from payrolls to revenue. Your company hasn’t innovated, your product hasn’t become more attractive or competitive, and your business is just as profitable as it was prior to the AI spending.
If revenue remains the same and costs (such as labor costs) go down, then profits increase.
This is very attractive to business owners; it's probably the most effective pitch for any kind of automation.
Spoken like an MBA. There are pitfalls in your way of thinking, but as they say—learn by doing.
[dead]