Learn and master Clay with Clay university

Welcome to the New Clay! Clay looks very different than it used to, and this course is a tour of how it works now: AI-first, built mostly by prompting, and then checked by you. Across the five core lessons we'll run one real play from search to action.

📣 Looking for the old Clay 101 course? That's moved to Legacy Clay 001.

🧭 The One Principle: Understand Your Outputs

Everything in this course is really about AI GTM. The platform is built for it, but the principle holds even if you run your GTM somewhere else, like Codex or Claude Code.

You have to be able to read and understand what your systems produce. GTM results aren't binary.

Say you ask an AI agent to build a system that finds work emails, and it finds 70% of your contacts. Is that great, or is it leaving half the list on the table? You can't tell unless you understand the waterfall logic, the AI prompts, the web research it falls back on, and how the whole system fits together.

That's what's different about this version of Clay 101:

• It's AI-first. Most of the building happens by prompting an agent through the Clay CLI, which is the most effective way to build today.
• You still learn every piece. We'll stop and look at what the CLI built, step by step, so you can troubleshoot it.

One exception for now: search doesn't run from the CLI yet, so the next lesson starts in Clay's own search.

🗂️ Audiences is Your Master Database

Open the left sidebar and you'll see two new entries: People and Companies. These are Audiences.

Think of Audiences as a master database of every person and company you've ever prospected, reached out to, or acted on. Instead of hunting through tables to find which contact landed in which campaign, everything lives in one record of truth that sits underneath your tables and workflows. It stays in sync with your CRM on the basic properties.

A simple way to remember this:

• Audiences are the records your old tables used to hold
• A segment is a slice of an audience you can act on

⚙️ Workflows: Logic First, Data Second

Workflows are the enrichment and logic layer built on top of Audiences. Open Workflows (still in beta) and start a new one, and there are two ways in:

• A blank chat, where you describe what you want built
• A node-based builder, if you'd rather build by hand

A workflow is one layer abstracted from a table: instead of putting your data first, it puts the logic of your columns first. You can still flip to a table view inside a workflow, so you get the best of both. And you get things tables made hard: code, proper branching, and conditional logic.

🧰 Everything Else That's New

Beyond Audiences and Workflows, Clay now has:

• Ads, for running ads on a specific segment
• A native Sequencer for email outreach
• An MCP for using Clay inside AI tools
• Functions for reusable logic
• Account Agents
• Claygents, Clay's AI research agents

We know, it’s a lot! Being overwhelmed is the most intimidating part. The good news is you don't have to memorize every feature.

What you need to remember is what's possible and what great looks like, because that's what lets you troubleshoot the first draft AI hands you. Clay ships fast, so a button may have moved by the time you watch this. The principles won't have.

🚀 What's Next

You now know the shape of the new Clay: Audiences as the database, Workflows as the logic on top, and AI as the way you build.

Next, we'll start the play in the new search experience, finding companies, people, and jobs.