24/7 Lead Qualification and Booking

Connect your website contact form to n8n via webhook. When someone submits the form, the data arrives in n8n immediately. Most form platforms (Typeform, Gravity Forms, HubSpot Forms, Webflow) support webhook output natively - find the "Webhooks" or "Integrations" section of your form settings and point it to your n8n webhook URL. After the webhook trigger, add an AI node to qualify the lead. Pass the full form submission data and use the qualification prompt below. The AI returns a qualification score (0 - 100), a status (qualified, not_qualified, needs_human_review), a summary of what the lead is looking for, and a suggested response. Branch the workflow based on the score. Leads above your threshold (start at 70) get a personalized response email generated and sent within two minutes. Leads below threshold go to your exception queue. Both paths create or update the CRM record and log the activity automatically.

You are a lead qualification specialist for a professional services firm. Your job is to evaluate inbound inquiries and determine whether they meet our qualification criteria.

Qualifying signals: {{your_qualifying_signals}}
Disqualifying signals: {{your_disqualifying_signals}}

Here is the lead submission: {{$json}}

Return ONLY a valid JSON object with these fields:

{
  "qualification_score": number between 0 and 100,
  "qualification_status": "qualified", "not_qualified", or "needs_human_review",
  "qualification_reason": "2-3 sentences explaining your assessment",
  "what_they_need": "Plain summary of what this lead is looking for",
  "suggested_response": "A complete, personalized email response that: (1) acknowledges what they said specifically, (2) answers the most likely initial question, (3) if qualified - includes a direct booking link placeholder [BOOKING_LINK], (4) sounds like a thoughtful person wrote it, not a template",
  "crmfields": {
    "lead_source": "web_form",
    "company": "extracted company name or empty string",
    "budget_signal": "high, medium, low, or unknown based on any signals in the submission",
    "urgency": "immediate, this_quarter, exploring, or unknown"
  }
}

Your intake email connects via a polling trigger in n8n - the same approach as Play One. Set polling to every 5 minutes. Add a filter to distinguish inbound lead inquiries from existing client emails (check whether the sender is already in your CRM as a contact; if they are, route to Play One's email logging workflow instead). For emails that don't match existing contacts, pass them to the same AI qualification node you built for the web form track. The qualification logic is identical - the AI assesses the email content against your criteria and generates a personalized response. The key difference in the email track: the AI needs to handle edge cases like forwarded introductions, people who email multiple times before getting a response, and inquiry emails that arrive as replies to automated confirmation emails from your booking system. Add handling logic for each of these - check the email subject line for "Re:" and cross-reference with your CRM before sending a first-contact response.

For the chat widget, you can build your own using Claude Code, ChatGPT Codex, or Gemini - these tools generate a functional widget in minutes from a plain-English description. The widget is a single JavaScript file you embed on your website. Tell the AI tool: "Build me a chat widget that connects to a webhook URL. When a user sends a message, POST it to [your n8n webhook URL] and display the response. Style it with a blue launch button in the bottom right corner." Once the widget exists, n8n handles the intelligence. When a visitor sends a message, n8n receives it via webhook and passes it to the AI with context about the ongoing conversation (use n8n's session memory or a simple key-value store to maintain conversation state). The AI asks your qualification questions naturally, answers questions about your services, and when the visitor is ready to book, n8n queries your calendar for available slots and passes them directly into the chat window. The full chat flow should handle: greeting the visitor, working through qualification questions naturally, answering common FAQs about your firm, offering booking slots, and gracefully routing to the exception queue if the conversation can't be resolved.

You are a professional intake specialist for {{firm_name}}, a {{firm_type}} firm. You're handling an inbound inquiry via live chat.

Your goals: (1) make the visitor feel heard, (2) understand what they need, (3) qualify them against our criteria, (4) guide them toward booking a call if they qualify.

Qualification criteria:
Qualifying signals: {{qualifying_signals}}
Disqualifying signals: {{disqualifying_signals}}

Conversation so far: {{conversation_history}}
Visitor's latest message: {{current_message}}

Guidelines:
- Sound like a knowledgeable, warm human - not a form or a phone tree
- Acknowledge what they said before asking your next question
- Ask one question at a time
- If they ask about pricing, give a range or offer to discuss on a call rather than dodging
- If they qualify, offer to book a time: "Based on what you've shared, it sounds like we'd be a good fit. Can I get you on the calendar with [partner name]?"
- If they don't qualify, be honest and helpful: suggest an alternative if you know of one

Return ONLY a JSON object:
{
  "response_text": "Your next message to the visitor",
  "conversation_status": "ongoing", "ready_to_book", "not_qualified", or "needs_human",
  "qualification_score": number 0-100
}

Create a Retell account (retellai.com) and configure a voice agent. The agent configuration defines: the voice profile (choose one that matches your firm's tone - professional, warm, direct), the greeting message, and the knowledge base it can draw from when answering questions about your firm. In n8n, create a webhook workflow that Retell calls during live conversations. Retell uses this webhook for three purposes: (1) CRM lookups at the start of a call - n8n checks whether the caller's number matches an existing contact, and returns their history so the agent doesn't treat a current client like a stranger, (2) calendar availability - when the conversation reaches the booking stage, Retell calls n8n for open time slots; n8n queries your calendar and returns them, (3) post-call logging - when the call ends, Retell sends the transcript and outcome to n8n, which creates or updates the CRM record and routes to the exception queue if the call couldn't be resolved. Configure your phone system to forward calls to Retell after hours and when the line is busy. Test the full flow by calling your own number: confirm the greeting sounds natural, the qualification questions flow conversationally, the booking process works correctly, and the post-call CRM update appears within 60 seconds of hanging up.

All three tracks feed the same exception queue. Create a dedicated Slack or Teams channel - #lead-exceptions or #intake-queue - with the owner named in the channel description. Exception queue items fall into three categories. Route each with the appropriate context: 1. Below-threshold leads: include the lead source, company name, what they're looking for, the AI's qualification score and reason, and the original submission text. The queue owner decides whether to respond manually or dismiss. 2. Conversation stuck or human requested: include the full conversation transcript, the channel it came in on, the visitor's name and contact information if captured, and the point where the conversation stalled. 3. Workflow errors: include which node failed, the input that caused the failure, and the lead's contact information so the queue owner can follow up manually. The owner should check the queue in the morning and after lunch. In the first two weeks, the queue will be busier than it will be long-term - every exception item is information about where your criteria need tightening or your prompts need adjustment.