KC screen-shared eight minutes of how he researches a high-value lead and replies. The bot saved it as a skill. Each morning it finds leads, scores them, drafts in his voice, and only pings the ones worth sending.
You already know how you research and reply to a good lead. You do not want to write that playbook from memory.
One taught skill. A morning pass of scored leads and drafts. You only see the ones worth sending.
Screen-share research plus the reply. Whole thing.
It watched. Do not write the SOP from memory.
Find. Score. Draft in your voice.
You send.
One loop. Clean before you add a seat.
What's good: show it once instead of writing the SOP. The catch: it drafts. You send. If the morning pile is slop, teach the pass again before you schedule it.
Watch me research a high-value lead and write the reply. Save that pass as a skill. Every morning: find new leads, score them, draft replies in my voice. Only ping me for the ones worth sending. I send. You do not send. If I correct you, update the skill. Do not start a second job until this loop is clean.
Rewritten from https://x.com/karanC_12/status/2089952478091751684
Write-ups follow public posts and official docs. No likes, views, or dashboard numbers. Open the source to check the original. About