Quick Start
Get your Hay AI assistant up and running in less than 10 minutes! Follow these simple steps to start automating your customer support.
Before You Begin #
Make sure you have:
- A Hay account (sign up at your organization's Hay dashboard)
- At least one support document or FAQ (we'll use this to train your agent)
- 10 minutes to complete the setup
Step 1: Choose Your Integrations #
Hay works where your team already does. You can connect integrations now or skip this and come back later.
- Click Integrations in the left sidebar
- Browse available integrations (Shopify, WhatsApp, Zendesk, etc.)
- Click on an integration you want to add
- Follow the simple connection steps
Popular integrations:
- Web Chat - Add a chat widget to your website
- WhatsApp - Handle WhatsApp Business messages
- Shopify - Access order and product information
- Zendesk - Sync with your help desk
Tip: You can start with just the Web Chat widget and add more integrations later!
Step 2: Create Your First Agent #
Your agent is your AI assistant. Give it a personality and purpose.
-
Go to Agents in the left sidebar
-
Click Create Agent
-
Fill in the basic information:
- Name: Something descriptive like "Customer Support Agent"
- Description: What this agent handles (e.g., "Handles order questions and product inquiries")
- Tone: Choose how your agent should communicate (Professional, Friendly, Casual, etc.)
-
(Optional) Add specific instructions:
- How to greet customers
- What topics to avoid
- When to escalate to a human
-
Click Create Agent
Example: "Support Agent" with a "Friendly" tone that "helps customers with orders, shipping, and product questions"
Step 3: Upload Training Documents #
Teach your agent by uploading your existing support materials.
-
Go to Documents in the left sidebar
-
Click Import Documents
-
Upload files or paste text:
- PDFs - Product manuals, policy documents
- Text files - FAQs, support scripts
- Direct text - Paste in common questions and answers
-
Wait a few seconds while Hay processes and learns from your documents
What makes good training material?
- ✅ Frequently Asked Questions
- ✅ Product documentation
- ✅ Return and refund policies
- ✅ Shipping information
- ✅ Common support scenarios
Tip: Start with your top 10-20 FAQs. You can always add more later!
Step 4: Create a Playbook (Optional but Recommended) #
Playbooks are step-by-step instructions for handling specific situations.
-
Go to Playbooks in the left sidebar
-
Click Create Playbook
-
Set up the basics:
- Title: "Welcome New Customers"
- Trigger: "greeting" or "hello"
- Instructions: What should happen when triggered
-
Write instructions in plain language:
Greet the customer warmly. Ask how you can help them today. If they ask about orders, check their order status. If they need help with products, provide relevant information. -
Click Save and Activate
Common playbook examples:
- Order status lookups
- Return and refund requests
- Product recommendations
Step 5: Test Your Agent #
Before going live with customers, test how your agent responds.
-
Go to Conversations in the left sidebar
-
Click + Conversation Playground
-
Start typing messages like a customer would:
- "Hi, I need help with my order"
- "What's your return policy?"
- "Do you ship internationally?"
-
Review the responses - do they sound right?
-
If needed, go back and:
- Add more training documents
- Adjust agent instructions
- Update your playbooks
Tip: Test with real questions your support team receives daily!
You're All Set! 🎉 #
Your Hay agent is now ready to start helping customers!
What Happens Next? #
- Automatic responses: Your agent handles incoming questions instantly
- Human escalation: Complex issues are automatically sent to your team
- Continuous learning: Your agent gets smarter with every conversation
Next Steps #
Now that you're up and running:
- Monitor the Queue - Check for conversations that need human attention
- Review Analytics - See how your agent is performing
- Add More Documents - Keep training your agent with new information
- Create More Playbooks - Automate more types of requests
Need help? Head to the Dashboard guide to learn about monitoring your conversations.