How to Set Up a GoHighLevel AI Agent for Automatic Lead Follow-Up (2026 Guide)
Step-by-step tutorial for setting up GoHighLevel's Conversation AI agent to automatically respond to leads, qualify them, and book appointments. From someone who builds these for clients.
TL;DR
GoHighLevel's Conversation AI lets you set up an AI agent that responds to leads via SMS, email, and web chat in under 60 seconds. Setup takes 2-4 hours: enable the AI Employee add-on, build a knowledge base, configure conversation flows, connect your calendar, and set up trigger workflows. This guide covers every step with exact settings and configuration tips.
GoHighLevel's Conversation AI is the fastest way to get an AI agent responding to your leads without writing code. I've set up these systems for real estate agents, insurance brokers, dental clinics, and service businesses — and the setup process is the same every time.
This guide walks through every step. If you want the bigger picture on why AI follow-up matters and how the pieces fit together, read the AI lead follow-up automation overview first. If you need convincing on why speed matters at all, the speed-to-lead statistics make the case with hard data.
This tutorial assumes you already have a GoHighLevel account. If you don't, sign up for the 14-day free trial first.
What GHL AI Agents Can (and Can't) Do
Before diving into setup, set your expectations:
What GHL AI agents handle well:
- Instant first response to new leads (SMS, email, web chat, WhatsApp)
- Answering FAQs about your business, services, and pricing
- Asking qualifying questions based on criteria you define
- Booking appointments on your GHL calendar
- Tagging and categorizing leads based on conversation
- Handing off to a human when needed
What they don't handle well (yet):
- Complex price negotiations or custom quotes
- Handling angry or escalated customers
- Multi-step sales conversations that require deep product knowledge
- Outbound cold calling (use Bland.ai or Vapi for that)
The sweet spot: let the AI handle the first 2-5 messages, qualify the lead, and either book an appointment or route to your team. Don't try to make it close deals.
Prerequisites Before You Start
Make sure you have these ready:
- [ ] GoHighLevel account (Agency or SaaS Pro plan)
- [ ] AI Employee add-on enabled ($49/month)
- [ ] At least one phone number set up in GHL (for SMS)
- [ ] A calendar configured with your available time slots
- [ ] Your FAQ document (15-20 common questions and answers)
- [ ] Your qualifying criteria written out (what makes a good lead vs. bad lead)
- [ ] Your business description (2-3 paragraphs covering what you do, who you serve, what makes you different)
Step-by-Step Setup
Step 1: Enable the AI Employee
- Go to Settings in your GHL sub-account
- Click AI Employee in the left sidebar
- Toggle Enable AI Employee to ON
- Select your plan tier (start with the base — you can upgrade later)
- Accept the terms
You'll see the AI Employee dashboard with options for Conversation AI, Reviews AI, Content AI, and Workflow AI. We're focused on Conversation AI.
Step 2: Create Your AI Agent
- Go to Conversation AI → AI Agents
- Click Create Agent
- Fill in the basics:
- Agent Name: Give it an internal name (e.g., "Lead Qualifier" or "Sarah — Front Desk AI")
- Agent Role: Select "Lead Qualification" or "Customer Support" depending on your primary use case
- Communication Channels: Enable SMS, Email, and Web Chat. Enable WhatsApp if you have it connected.
Step 3: Build the Knowledge Base
This is the most important step. Your AI agent's quality is directly tied to the quality of information you give it.
- Go to Knowledge Base within your AI Agent settings
- Click Add Source and choose from:
- Website URL — paste your website and GHL will crawl and extract content
- Document Upload — upload a PDF or text file with your FAQ and business info
- Manual Entry — type or paste content directly
What to include in your knowledge base:
| Category | What to Write | Example | |---|---|---| | Business Overview | What you do, who you serve, how long you've been in business | "We're a dental practice in Miami serving families since 2015. We specialize in cosmetic dentistry and Invisalign." | | Services | Each service with a brief description and price range | "Teeth whitening — in-office treatment, 1 hour, $350-500. Invisalign — 12-18 month treatment, $3,500-6,000." | | FAQ | The 15-20 questions leads ask most | "Do you accept insurance? Yes, we accept Delta Dental, Cigna, and Aetna." | | Differentiators | What makes you different from competitors | "We're one of three Invisalign Diamond providers in South Florida, which means we've completed 800+ cases." | | Qualifying Criteria | What makes a lead a good fit | "Qualified lead: lives within 30 miles, has dental insurance or budget for self-pay, looking for treatment within 3 months." | | Disqualifying Criteria | What makes a lead NOT a fit | "Not a fit: looking for emergency extraction only (we don't do those), lives outside service area, only wants a free consultation with no intent to proceed." |
Pro tip: Write the knowledge base in the same voice you'd use talking to a lead on the phone. The AI will mirror the tone of the content you give it. If your knowledge base is stiff and corporate, the AI will sound stiff and corporate.
Step 4: Configure the Conversation Flow
- Go to Conversation Settings in your AI Agent
- Set the System Prompt — this is the master instruction that tells the AI how to behave. Here's a template:
You are [Agent Name], a friendly and professional assistant for [Business Name].
Your job is to:
1. Welcome new leads and make them feel heard
2. Answer questions about our services using the knowledge base
3. Ask qualifying questions to understand their needs
4. Book qualified leads on our calendar
5. Escalate to a team member if you can't help
Rules:
- Be conversational, not robotic. Use short sentences.
- Never make up information. If you don't know, say "Let me connect you with [rep name] who can help with that."
- Never quote exact prices unless they're in the knowledge base.
- Always try to move the conversation toward booking an appointment.
- If someone seems frustrated or asks to speak with a human, immediately offer to connect them.
- Keep messages under 3 sentences. Nobody reads walls of text.
- Set Max Messages Before Escalation — I recommend 8-10. If the AI hasn't booked or escalated by then, something is wrong.
- Configure Escalation Rules:
- Trigger phrases: "speak to someone," "talk to a human," "manager," "complaint"
- Unrecognized questions: after 2 consecutive "I don't know" responses
- Set the escalation action: notify a specific team member via SMS + email
Step 5: Connect Your Calendar
- Go to Calendars in your GHL account
- Make sure you have a calendar set up with:
- Available days and hours
- Meeting duration (15, 30, or 60 minutes)
- Buffer time between meetings (at least 15 minutes)
- Meeting type name (e.g., "Free Consultation," "Discovery Call")
- Go back to your AI Agent → Appointment Booking
- Select the calendar the AI should use
- Configure the booking message template:
Great! I have availability on [date] at [time]. Would that work for you?
If not, I can suggest a few other times.
- Enable Confirmation Messages — the AI will send a text confirming the booking with date, time, and any prep instructions you define.
Step 6: Build the Trigger Workflow
The workflow tells GHL when to activate the AI agent.
- Go to Automations → Workflows
- Click Create Workflow → Start from scratch
- Add a Trigger:
- Trigger type: Contact Created (fires when a new lead enters the CRM)
- Optional filters: filter by lead source (e.g., only Meta Ads leads) or tag
- Add a Wait step: 10 seconds (gives the CRM time to populate contact data)
- Add an AI Agent action:
- Select your AI agent
- Choose the communication channel (SMS is best for first response)
- Set the initial message: reference what the lead asked about if possible
- Add Conditional branches after the AI conversation:
- If appointment booked: tag as "qualified," add to sales pipeline, notify rep
- If not qualified: tag as "nurture," add to email drip sequence
- If escalated: tag as "needs-human," send internal notification with conversation transcript
Your workflow should look like this:
[New Contact Created]
↓ (wait 10s)
[AI Agent: Send SMS]
↓ (conversation happens)
[Check: Appointment Booked?]
├── YES → Tag "qualified" → Add to pipeline → Notify rep
├── NO (nurture) → Tag "nurture" → Add to email sequence
└── ESCALATED → Tag "needs-human" → Alert team + send transcript
- Activate the workflow — toggle it to "Published"
Step 7: Set Up Multi-Channel Triggers
One workflow isn't enough. You need separate triggers for each lead source:
| Lead Source | Trigger | Channel | Notes | |---|---|---|---| | Meta Ads Lead Form | Form Submitted (Meta integration) | SMS | Include ad name in first message for context | | Website Form | Form Submitted (GHL form) | SMS + Email | Use both channels for higher contact rate | | Web Chat | Chat Widget Message | Web Chat | AI responds in the chat widget directly | | WhatsApp | WhatsApp Message Received | WhatsApp | Great for LATAM markets | | Instagram DM | Instagram DM Received | Instagram DM | Keep responses short and casual | | Google Business | GBP Message | SMS | Respond then move to SMS for continuity | | Manual/Referral | Contact Created (manual) | SMS | Rep creates contact, AI follows up |
Step 8: Test Everything
Do not go live without testing. Run through these scenarios:
- Submit a test lead through Meta Ads (use a test campaign or manual form submission). Verify the AI responds within 60 seconds via SMS.
- Answer the AI's questions as a qualified lead. Verify it offers appointment times and books correctly.
- Answer as a non-qualified lead (wrong budget, wrong location). Verify it tags correctly and enters nurture.
- Ask a question the AI can't answer. Verify it escalates properly and your team gets notified.
- Test after hours. Submit a lead at 10 PM. The AI should respond identically — no "our office is currently closed" messages.
- Test the booking confirmation. After booking, verify the lead gets a confirmation SMS and the appointment appears in your calendar.
- Check the CRM. Every conversation should be logged in the contact record with full transcript.
Configuration Tips From Real Deployments
After building 30+ of these systems, here's what I've learned:
Start with SMS only. Don't enable every channel on day one. Get SMS working perfectly, then add email, then web chat, then WhatsApp. Each channel has its own quirks.
Keep the first message under 160 characters. SMS has length limits, and longer messages split into multiple texts that arrive out of order. The first message should be short: "Hey [name], thanks for reaching out about [service]. Quick question — [qualifying question]?"
Don't over-qualify. Two qualifying questions is enough for the AI. Three is the max. More than that and leads drop off. Get the basics (what they need, when they need it) and book the call. Your sales team handles the rest.
Review conversations weekly. For the first 4 weeks, read every AI conversation. You'll find patterns: questions the AI handles poorly, responses that confuse leads, qualification criteria that's too strict or too loose. Update the knowledge base and system prompt based on what you find.
Set up a "human takeover" button. In GHL, you can manually jump into any AI conversation from the Conversations tab. Train your team to check active conversations and take over when they see a high-value lead or a conversation that's going sideways.
When to Build It Yourself vs. Hire an Expert
Build it yourself if:
- You're comfortable navigating GHL's interface
- You have 4-8 hours to dedicate to setup and testing
- Your business has straightforward lead qualification (1-2 criteria)
- You want full control over the system
Hire an expert if:
- You need multi-channel setup (SMS + WhatsApp + web chat + email)
- Your qualification process is complex (multiple services, locations, or team members)
- You want custom integrations (external calendar, Slack notifications, CRM sync)
- You'd rather spend your time on sales and let someone else handle the tech
- You need bilingual support (English + Spanish) configured properly
If you're in the "hire an expert" category, that's exactly what we do. We build complete AI follow-up systems on GoHighLevel — from knowledge base to workflows to testing — so our clients can focus on closing deals instead of configuring software. Book a call to see if it's a fit.
The difference between a business that responds in 4 hours and one that responds in 30 seconds isn't talent or budget. It's systems. GoHighLevel gives you the tools. This guide gives you the blueprint. The only thing left is building it.
If you're not sure whether AI follow-up is right for your business — or which setup makes sense for your situation — take the AI readiness quiz. It takes 2 minutes and shows you exactly where to start.
FAQ
How much does GoHighLevel AI agent cost?
GoHighLevel starts at $97/month for the base plan. The AI Employee (Conversation AI) is an add-on at $49/month, which includes a set number of AI interactions. Additional interactions are billed at roughly $0.02-0.05 per message depending on volume. Total cost for most small businesses: $150-200/month.
Does GoHighLevel AI work with WhatsApp?
Yes. GHL supports WhatsApp Business API integration. You can connect your WhatsApp number and have the AI agent respond to incoming WhatsApp messages using the same conversation flows. This is especially valuable for LATAM markets where WhatsApp is the primary communication channel.
Can the GoHighLevel AI agent book appointments?
Yes. The AI agent integrates with GHL's built-in calendar system. You configure available time slots, meeting types, and booking rules. When a lead is qualified, the AI offers available times and books directly — no human needed. The appointment shows up in your calendar with the lead's contact info and conversation summary.
Is GoHighLevel better than HubSpot or Salesforce for AI lead follow-up?
For small-to-mid businesses doing $3K-50K/month in revenue, GHL is typically the best value because the CRM, AI agent, workflow automation, and calendar are all built in. HubSpot and Salesforce require third-party AI integrations that add cost and complexity. For enterprise with existing Salesforce deployments, adding AI through their ecosystem makes more sense.
What if I don't want to set this up myself?
Many businesses prefer done-for-you setup. An experienced GHL automation specialist can build your complete AI follow-up system — including knowledge base, conversation flows, workflows, and testing — in 1-2 weeks. This typically costs $1,500-5,000 depending on complexity.
Related Articles
Ready to level up your marketing?
Book a free discovery call and see how AI can transform your results.