1. Collect data
    2. ???
    3. Profit

In this case, I doubt having that data will do much; humans are already natural optimizers. Really, the things workplaces should be optimized for are unlikely to be easily measurable, so at best these systems will optimize for the wrong thing.

And really, the reason workplaces aren't particularly optimal now is that most decisions are made not for maximizing business value, but for maximizing the power and comfort of those high up. As an example, when businesses relocate their headquarters, it is generally to move closer to the CEO's home. Or we can look at all the decisions made to optimize a visible metric, like cost, to the harm of invisible ones. E.g., the open office plan, which is cheap and lets managers supervise by looking out their office window, but significantly harms productivity and employee happiness.