Today a fellow human (hello!) may still be processing your mortgage application but probably not your mobile phone plan application. Phone companies have made a calculation and decided that a poor approval for a mobile phone plan isn't worth the cost of someone reviewing each application. Mortgage companies have typically calculated differently and you spend some time discussing your finances with a mortgage officer.
Every business is going to be looking at the speed vs. cost vs. accuracy equation through a new lens as AI continues to progress over the coming years.
"Fast. Good. Cheap. Pick two."
In one sense, a business is the sum of its processes. Those can be clearly defined or very ambiguous. Businesses hire people and use software to run those processes.
Cost of task = Direct Cost (Labor, Compute, etc.) + (Cycle Time * Opportunity Cost) + (Quality Risk * Impact)
Cost of Process = SUM(Cost of each task)
Operating Expenses = SUM(Cost of each process)
Generally, the more ambiguous a task, the higher the cost for labor thus the drive for building standard operating procedures and breaking larger processes down into smaller, more well-defined, steps that can be automated or delegated to lower cost labor.1
The more a business values accuracy (quality risk), the more they will pay for labor. Ditto for speed.
For knowledge work (anything that involves processing data), businesses make decisions with this equation every day. Part of what will drive future data demands will be businesses trying to better quantify what those costs are. AI is going to draw extra attention to this.
When AI reduces Direct Costs and Cycle Time to fractions of their former selves, the question becomes Quality Risk. How much does a negative (inaccurate) output on that task cost the business (impact)? Businesses that are able to answer that question, and align their investments and operations with that, are going to be the ones that will continue to succeed.2
Those that don't are going to fall behind as Cycle Time keeps decreasing for their competitors. The speed of business will continue to increase. Those successfully incorporating AI will move at a higher rate, make more moves, and ultimately book more revenue.
The innovation is happening. I'm enjoying speculating how certain industries will be transformed, and entirely new industries spring up, as the cost inputs in this equation change.3 As a whole, we're going to need less of us humans for each individual process. AI doesn't change the equation. It will level the playing field operationally. The winners will be the companies that redirect their people's creativity towards designing new processes that create and capture new opportunities.
Pay seems to be roughly (at least directionally) correlated to the number of inputs that go into the output that the position is judged on.
I am generalizing and considering innovation an operating process like any other.
A couple of assumptions I made in this are: people will cost more than AI and people are more accurate than AI. I expect the former will hold true. The latter is already changing and those are going to be the first jobs at risk (outside creative).