There's also some weird side effects on how you set your prices.
When I was doing security consulting, I often left clients with a list of things they need to do, and things they need to frequently check, for security purposes - things like, maintaining firewalls between departments instead of just at the perimeter, checking for new exploits against their routers and switches on a regular basis, keeping firmware updated, not putting random Internet-of-Targets crap on the same networks as valuable servers, disabling remote-management interfaces where not needed, resetting default passwords on machinery they had never even realized HAD remote management features, disabling USB and wi-fi on the motherboards of everything in the security enclave, NOT having the database of all customer information unencrypted on a server available to all employees, etc etc etc...
The likelihood of people following these instructions was directly proportional to the amount of money I was charging them. Some irrational psychological effect says that something which was expensive is more valuable.
I eventually did the math and realized, my customers actually got more for their money if I charged them triple the rate I had been charging. Providing exactly the same service for triple the price got them to actually follow instructions which saved them from expenses even greater.
But the weird fakeouts of irrational pricing effects don't stop there. If I went to one of my old customers and gave them a *BILL* that gave the tripled price and then a huge discount for our longstanding business relationship totalling out to about the same amount I had been charging them.... that was almost as effective for purposes of security - in terms of getting them to follow instructions - as actually charging them triple.
Artificial Intelligence consulting is also irrationally priced, but it's a different set of irrationalities. People aren't usually left with instructions that require them to proactively maintain and monitor things that are inconvenient or that they would otherwise ignore, so the "won't follow instructions because they were cheap" effect isn't as dire for their return-on-investment. However, there are some subsets and specialties in this business where people overstate what they can achieve (and charge) specifically *because* they lack the knowledge to know why they can't or what makes it hard. This also has the effect of inversely correlating cost and benefit for the customer within those particular subsets and specialties.
Doing AI chassis mods and repairs? I don't know that business; I don't think we have a real equivalent. But the question of how to price your services and whether a higher or lower price provides more value for customers is far from trivial.