Sometimes its very important to say what features are absolutely necessary or required and what features are "Nice to have". Then a brainstorming or review can be done and that's it. Just because the tool can support and people can play with, the system should not be compexed. Many projects fail because they try to get everything on board rather than concentrating on what is required.
Adding a robot and multiple algorithms to a vision system makes it look like the AOI systems I'm used to see in PCB assembly. They are pretty expensive, being off-the-shelf. So I imagine a custom solution would be cost prohibitive.