Define Chatbot Logic (10 mins)

Use Durable AI: “Create a chatbot for my SaaS site that answers FAQs about pricing, features, and onboarding. Redirect to Calendly after 3 messages.”

Output: Embed code + basic UI.

Customize Frontend (15 mins)

Prompt v0.dev: “Generate a React chat widget with purple gradients and a ‘Talk to AI’ button.” → Copy/paste code or import your code into Cursor (see step 4).

Upload Data:

Go to Lindy’s “Knowledge Base”:

• Upload PDFs/CSVs of: -- Past support tickets (e.g., “How to reset passwords?”). -- API documentation. -- Refund policy.

Train Model: Here is an example prompt: "Analyze my support tickets and create responses for:

• Login issues (link to password reset page).
• Feature requests (reply with ‘Coming in Q3!’).
• Refund requests (direct to support@company.com).
• Use empathetic language and avoid markdown.”

Integrate with Durable:

• Copy Lindy’s API endpoint → Paste into Durable’s “Custom Knowledge” tab.

Optimize with AI Code Assist (15 mins)

Use Cursor Pro: “Add rate-limiting to the OpenAI API to keep costs under $50/month.” → Auto-insert code snippets.