Ground Your AI: Build an Assistant That Knows Your Documents

An AI that makes things up with confidence is worse than one that admits it doesn’t know.

Most generative AI tools guess when they don't have the facts. In this hands-on session, you’ll learn how to ground AI in your own real documents so every answer can be traced directly back to verified source material.

We’ll demystify RAG (Retrieval-Augmented Generation) without the technical jargon and explore why clean, well-organized files make or break your AI's accuracy. You’ll build a no-code assistant using real documents and test whether its answers stay tied to what you gave it.

🛠 What We’re Doing

• Build a No-Code Assistant: Upload real documents and configure an assistant that speaks strictly from your provided facts.

• Audit for Hallucinations: Test your assistant's boundaries and learn how to catch it when it tries to make things up.

• Document Hygiene 101: Discover why messy PDFs confuse AI and how to structure your files for maximum clarity.

No coding background needed. This is about literacy and judgment, not development mechanics.

New? Jump in anytime. Every class starts with a quick recap, so you won't be lost.

Free. Join in person or on Zoom.

Part of the AI Architect Fellowship, a free program from 3percentclub for underrepresented builders in NYC.

Joining virtually? Live stream: https://us06web.zoom.us/j/83454908023

Applying to the fellowship gets you:

• Access to our Google Classroom for labs and coursework

• Mentorship from working AI practitioners

• Resume workshops

• Our Discord community

• A shot at the October hackathon and Demo Night

Sign up: https://forms.fillout.com/t/p52mmdNs8jus