I don’t know if I’m misunderstanding how you’ve put this, but I’ve only ever understood dogfooding to mean that you use your own product to find the bugs, improve the workflow, and encourage feature suggestions.
Dog-fooding is not a reward for making the best product (though that makes life easier at work), it’s a commitment to improving your own product, and showing that you are willing to live with your own product in order to make it better for your customers.
So yes, dogfooding is there to get immediate feedback on technical flaws of an otherwise desirable product. Dogfooding is not there to fix a product--market mismatch. That needs to happen at a different level of development.
I've only ever understood dogfooding to mean this as well, but I like GP's point. If you're forcing yourself to use the product to find bugs instead of using the product because it's a great product, then you have problems.
You wouldn't want to eat dog food that wasn't prepared in sanitary conditions, so if you worked at a dog food plant and knew you'd be eating something that closed through that dirty grate, you'd clean it. You wouldn't want a program on your PC that regularly consumes an entire core while idle or that crashes when active, so you'd fix those issues if it was your software.
People naturally want to fix problems that affect them and add features that would improve their daily use of a product. At a small company or startup, where people can do this, you should use your own products as much as possible.
But if you are a Microsoft/GitHub employee with no access to the Teams code, not even access to a human being who works on Teams, there's no point. Use whatever meeting software lets you do your job best.
(It doesn't have to be perfect, or course, just do something better than its competitors to be attractive to use. I struggle to think of a single redeeming feature of Teams.)