@ayo_deshina That is the kind of bug that only shows up when someone listens to the actual calls, because the agent sounds just as confident reading 1 AM as 9 AM. Do you now run a live test call for every new client setup?
@PatSchaber Agreed, and the AI SDR can still take the top-of-funnel volume so the new grad spends their hours on real conversations. Which part of ramp do you see an AI coach shorten most?
@JakeDSBD24@CarlosFeIru Agreed, an exceptions list keeps that hour honest. The hard part is defining odd: a confused or annoyed reply from the prospect is an obvious one. Which signals do you put on the list first?
@harshjndigital Well said, and it gives the human a clear job too. If they read every email before it goes out, their edits are the best feedback the AI will ever get.
Our AI voice agent got better because a person listens to its live calls every day and ships fixes the same day. Budget human QA for AI the way you would for a new rep.
If you can't tell which outcomes your AI agent touched, you can't answer "is the AI working?" Tag every AI action in your system of record when it happens, not in a side spreadsheet afterwards.
If I hit 10,000 followers and officially crack this algo by the end of 2025…
I’m taking ONE trader under my wing.
Full mentorship. Everything I know. No fluff.
To enter:
• Follow me
• Like this post
• Repost it
❤️♻️🐦⬛