The most useful minute of a sales call turned out to be the one after the customer hangs up
We build AI phone agents. Over the past year we also added real-time assist for human-to-human calls: the AI listens in, suggests answers, surfaces FAQs, can fire off an SMS…
By Yusuf Ahmed
We build AI phone agents. Over the past year we also added real-time assist for human-to-human calls: the AI listens in, suggests answers, surfaces FAQs, can fire off an SMS while you're still talking. We also had post-call analysis: summaries, transcripts, insights. Looked great. The problem: it was built to understand calls, not to act on them. If I wanted something done after a call, I had to open Claude or ChatGPT and say "look at this call and do X". That extra step mostly never happened. Busy day, next call, forgotten. In a brainstorming session a colleague asked: what if, when the customer hangs up, the AI stays on the line and asks you what you want done? We tried it on onboarding calls (customers in their first 30 days). Customer hangs up, my side stays open, AI asks "anything to add?" and I just talk: "Find detailed instructions on X and send them …
It is part of a broader wave of investment and competition across the AI industry, as companies race to turn research progress into shipped products.
