So the question we kept having: why does all this intelligence only show up after the call ends?
We tried the obvious things. LLM in a side tab doesn't work, you can't always type and listen at the same time... and it knows nothing about your call anyway. Then I built the better version, an AI commenting on the call live. But we noticed that the real importance is minimizing a cognitive load of users by minimizing interruptions during the live call.
So almost all the work went into making it talk less, think more. This is how we built Heyalo (https://heyalo.ai). It listens and reasons through the whole call, speaks when something's worth a mention.
ungreased0675•57m ago