I mentioned PSI before. And I’m pretty sure I will again. Microsoft’s Platform for Situated Intelligence has a lot of potential. I was reminded about it today with this tweet:
From the PSI lab, they released a great demo that I’ve spent some time playing with.

With PSI, you can access all the streams of a call, and then modify those streams before sending data back down through a call. I took a while to get started last time I build a POC, but with more use cases and data, I should get even better at it.
Once I do, I’ll be pushing to show examples of applying deep intelligence during Teams calls. This should be fun!