[MattGaiser]

Probably depends on what exactly the product is.

For example, many consulting firms do not really sell technology. They sell handholding for organizations scared/unable to figure out technology and they quite often screw up the actual technology itself without any real consequence. There are plenty of companies where whether something actually works is not a key determinant of sales/revenue.

[rxm233]

In consulting "AI" firms, do you think engineering is more necessary than the actual AI part? To me, it seems that in those situations, knowing when and where to apply certain models could actually solve a lot of business problems for a client.