Om Asnani

@omasnani

When Analysts Meet Agents: Teaching Semantic Workflows Without Tool Soup

Submitted Sep 26, 2026

Description

Most “AI for analysts” demos stop at a clever chat window. The hard part starts when a team has five tools, three prompt libraries, and still no shared definition of a good answer.

I run AI learning programs at Outskill (GrowthSchool founding team) for people who live in sheets, BI tools, and CRM notes, not model cards. This meetup talk is a field report from teaching semantic and agentic workflows to non-research users: what made a workflow stick for a week, what collapsed into copy-paste theater, and how we score whether work actually changed.

Takeaways

You will leave with:

  1. A one-page brief, retrieve, draft, human verify pattern for analyst tasks
  2. Three failure modes I keep seeing: tool soup, eval theater, and demos that never become habits
  3. A simple rubric community and data leads can use before they buy another AI seat

Not a product pitch. Operator stories from teaching at scale. Happy to do a live walkthrough of a small workflow if the room wants a demo format.

Who is this for?

Analysts, data leads, operators, and community builders teaching or adopting AI workflows in sheets, BI tools, and CRM systems.

Bio

Om Asnani is Head of Community and Learning Experiences at Outskill. Founding GrowthSchool team (5 to ~350). Builds practical GenAI programs for engineers and operators, community systems at Discord/Telegram scale, and agentic workflows people keep using after the workshop ends.

Company: Outskill
City: Bengaluru
LinkedIn: https://www.linkedin.com/in/omasnani/
Format: Talk / case study (demo optional)
Name: Omprakash Asnani / Om Asnani
Email: omasnani@gmail.com

Comments

{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hosted by

Jumpstart better data engineering and AI futures