Quickstart

Get Floe running on your product in about 15 minutes. By the end, you'll have an AI agent that understands your product well enough to demo it.

Prerequisites

You'll need access to the Floe dashboard. If you don't have an account yet, talk to the founders to get set up.

Step 1: Create a site

A site represents one product or application that Floe will learn about.

  1. Go to Sites in the dashboard
  2. Click Add Site
  3. Enter your product's URL (e.g., https://app.yourproduct.com)
  4. Give it a name (e.g., "YourProduct")

The site is the container for everything Floe learns about your product: content, screens, and guided plans.

Step 2: Add content via ingestion

Floe needs to read your product's documentation and knowledge base to understand what it does.

  1. Go to your site's Ingestion tab
  2. Click Add Config
  3. Choose a source type:
    • Documentation: Your product docs site (e.g., https://docs.yourproduct.com)
    • Knowledge Base: Help center articles
    • Video: YouTube or Loom product walkthroughs
  4. Enter the URL and save

Floe's ingestion pipeline will read and index all the content automatically. This typically takes 2-5 minutes depending on the amount of content.

Step 3: Run exploration

Exploration is where Floe navigates your product to understand the UI.

  1. Go to the Explorer tab
  2. Click Start Exploration
  3. Floe navigates your product and maps the interface

After exploration completes, you'll see what Floe discovered under the Explorer tab. This is what the AI uses to navigate your product during demos and onboarding.

Step 4: Generate capabilities

Capabilities describe what users can do on your product. Floe generates these by combining your content with the screens it discovered.

  1. Go to the Capabilities tab
  2. Click Generate
  3. Floe analyzes your content and screens, then produces a list of capabilities like "Create a new survey", "Invite team members", "Configure notification settings"

Each capability describes what users can do and how to get there. The demo agent uses these to know what to show and when. See how it works for the full picture.

Step 5: See it in action

Once capabilities are generated, your demo agent is ready.

  1. Go to Demo Sessions
  2. Click Start Demo
  3. The AI agent opens a real browser with your product, greets the prospect with voice, and walks through your product based on the generated capabilities

The agent adapts to what the user asks about. If they say "show me reporting," the agent navigates to the reporting section and explains what it learned from your docs and product screens. This is demo automation that compresses time to value.

What's next

FAQ

How long does the full setup take? About 15 minutes for the basics. Ingestion takes 2-5 minutes, exploration takes 5-10 minutes, and capability generation is near-instant.

Can I skip steps? Each step builds on the previous one. The agent needs content (ingestion) and product understanding (exploration) before it can generate useful capabilities. Skipping steps means a less capable agent.

What if I don't have documentation? You can add other content types: YouTube walkthroughs, Loom recordings, knowledge base articles. Any content that describes your product works. You can also start with just exploration and add content later.