A practical workshop for AI to Grow
AI Agents at Work: From Meeting to Verified Outcome
In one complete case, Andrew Tomin will show how an AI agent finds decisions and gaps in a meeting transcript, builds a working brief, and creates a prototype after human review.
What we will cover
One working scenario for different roles
The meeting is over. Now someone has to preserve the decisions, ask the right questions, and turn the conversation into a useful result. We will show that whole process without touring dozens of tools.
00-05
Chat versus agent
A simple look at goals, context, files, tools, agent actions, and human review.
05-12
Meeting transcript: decisions and gaps
Give the agent a transcript and identify decisions, contradictions, commitments, and missing information.
12-20
Clarification and working brief
A person answers the important questions, then the agent captures the goal, context, constraints, and completion criteria.
20-25
Human Check before building
Check the brief against the source, separate facts from assumptions, remove ambiguity, and only then give the agent its next step.
25-35
Prototype from the approved brief
The agent builds a clickable prototype, and we test whether it solves the original task on a working example.
35-40
Transfer to your role and participant pack
See how the same process applies to a product task, project plan, marketing brief, or operations workflow.
Complete workflow
Meeting transcript → decisions and gaps → human clarification → working brief → Human Check → prototype → outcome validation.
Next step
Build this workflow for your own task
In AI Sandbox, you bring a real work task and take it from context and a working brief to a verified tool. Group 7 starts on November 17 and is taught in Russian.