I think the 'proper' approach is to use whatever is economically viable to get to the desired result as quickly as possible.

Ultimately the market will only ask two things: Does it work? And, how much does it cost?

No-one cares that "Ah but ours only uses vision".