Grok Bot sales workflow demo shows where Jev's decision layer fits
A breakdown of SpaceXAI's Grok Bot sales automation workshop, arguing that bots guessing decisions in Slack prose should hand off to Jev, which returns typed options with probabilities.
SpaceXAI shipped a second free Grok Bot workshop 55 minutes of a sales org built out of bots: 02:27 - the pitch is work finishing overnight, on calls, while you walk the dog. it feels like handing real work to a colleague you trust 08:05 - you give the prospecting bot five accounts and five prospects and it goes off to do the research 18:41 - the first version gave you a link instead of watching the 45-minute podcast for 50 contacts 21:11 - you type "i need to create a webinar on grokbot, can you help me" and it assembles the team for you 30:45 - they made a slack channel where the bots complain. the top complaint: you told me to do the research, i don't have access So every "should i stop here" is a paragraph of prose in a slack thread Whether to escalate, retry, or quietly drop the task, the model is guessing with adjectives Jev returns typed options plus a probability, $0.042 per million input tokens, and never writes a sentence LLM creates the work -> Jev decides what happens next -> code enforces it Save this before your bot asks for help in slack
This post quotes another post: “Jev turns routing, scoring and verification into a decision layer”