The GoHighLevel AI Decision Maker is an intelligent workflow branch that uses AI to analyze contact data, conversation history, and custom fields — then routes leads to different paths automatically based on what the AI determines about their intent, readiness, and fit. Instead of rigid if/else conditions (“if tag = X, do Y”), the AI Decision Maker understands context and makes nuanced routing decisions.
Launched in the 2025 GHL platform update, it’s one of the most powerful additions to the workflow builder. This tutorial shows exactly how to set it up and what it does best.
AI & Business Automation Courses Learn AI automation with hands-on coursesLearn more →
AI Decision Maker vs Standard Workflow Conditions
Standard workflow conditions in GHL use binary logic: if a field equals a value, go one way; otherwise, go another. This works well for simple routing but breaks down with nuanced scenarios:
“Is this lead ready to buy or just researching?”
“Did this prospect’s SMS reply show a positive or negative sentiment?”
“Based on everything we know about this contact, what stage of the buyer journey are they in?”
The AI Decision Maker handles all of these. It reads the full context — previous messages, custom field values, tags, activity history — and makes a judgment call, similar to how a skilled sales rep would.
Step-by-Step: Setting Up the AI Decision Maker in GHL
Step 1: Open Your Workflow
In your sub-account, go to Automation → Workflows. Open an existing workflow or create a new one. The AI Decision Maker works best placed after an initial trigger and some data collection — for example, after a lead fills out a form or after a conversation AI chat session.
Step 2: Add the AI Decision Maker Action
Click the + button to add a new action. In the action search, type “AI” and select AI Decision Maker. You’ll see a configuration panel open.
Step 3: Write Your Decision Prompt
This is the most important step. Write a clear prompt describing what decision you want the AI to make. Be specific:
Weak prompt: “Is this lead interested?”
Strong prompt: “Based on this contact’s form submission, conversation history, and any custom field data, determine if they are: (A) Ready to book a call now — they’ve expressed urgency and meet our criteria, (B) Interested but need more nurturing — they’re a fit but haven’t expressed urgency, or (C) Not a fit — they don’t match our target client profile.”
You can also reference specific custom fields: “The contact’s budget field shows {{contact.custom_field.budget}}. Consider this in your decision.”
Step 4: Define the Outcome Branches
For each possible outcome in your prompt (A, B, C above), create a corresponding branch in the workflow. The AI Decision Maker will automatically route to the matching branch based on its analysis. You can have 2-5 branches — keep it manageable.
Step 5: Add Context Data
Under Context in the AI Decision Maker settings, specify what data the AI should consider:
Last SMS/email message content
Specific custom field values
Contact tags
Pipeline stage
Conversation summary
The more relevant context you provide, the more accurate the decisions. Don’t include irrelevant data — it confuses the AI.
Step 6: Test with Sample Contacts
Before activating, use the Test Workflow feature. Run 5-10 contacts through the workflow manually and verify the AI is routing them correctly. Adjust your prompt if the decisions don’t match your expectations.
Best Use Cases for the AI Decision Maker
Lead Qualification Routing
After a Facebook Lead Ad form submission, use the AI Decision Maker to route: qualified leads (match your ICP) → immediate booking sequence; semi-qualified leads → 7-day nurture; unqualified leads → low-touch email only. This replaces manual lead scoring that most teams never do consistently.
Conversation Sentiment Analysis
After a Conversation AI chat or SMS exchange, run the AI Decision Maker to assess sentiment: positive (ready to move forward) → sales team notification; neutral (needs more info) → send case study; negative (objection or complaint) → escalate to human immediately.
Re-engagement Campaign Routing
For a database of old leads (90+ days inactive), use the AI Decision Maker to assess each contact based on their original form data and any past activity: “Is this contact worth a direct phone outreach, an email campaign, or should they be archived?” Saves your sales team time by only surfacing the contacts worth calling.
Post-Appointment Follow-Up
After an appointment is marked complete, the AI Decision Maker analyzes the call notes (added via custom field) and routes: deal likely (positive notes) → send contract/proposal; deal uncertain → schedule follow-up call; deal lost → add to long-term nurture.
AI Decision Maker vs Conversation AI vs AI Voice Agent
Feature
AI Decision Maker
Conversation AI
AI Voice Agent
Primary function
Route contacts in workflows
Chat with leads via SMS/web
Talk to leads via phone
Input
Contact data + custom fields
Real-time text messages
Real-time voice calls
Output
Workflow branch selection
SMS/chat replies
Spoken conversation
Human interaction
None (backend)
Yes (direct with lead)
Yes (direct with lead)
Tips for Writing Better AI Decision Maker Prompts
Be specific about criteria — vague prompts produce inconsistent routing. Define exactly what “qualified” means for your business.
Use numbered outcomes — label each possible decision (1, 2, 3) and match them to workflow branches with the same names.
Include fallback logic — always have a “catch-all” branch for when the AI is uncertain.
Test with edge cases — run unusual contacts through and see if the AI handles them sensibly.
Iterate weekly — review routing decisions weekly for the first month and refine your prompt based on what’s wrong.
Final Thoughts
The GoHighLevel AI Decision Maker is a genuinely intelligent workflow tool that moves GHL automation beyond simple if/else logic. For agencies with complex lead qualification requirements or businesses that want to automate nuanced decisions without hiring an analyst, it’s one of the highest-ROI features in GHL’s 2026 platform. Start with one use case (lead qualification routing), validate it, then expand to other workflows.
⚡ Quick Summary
The GoHighLevel AI Decision Maker routes leads through your workflow using AI judgment — not just if/else rules. You write a prompt defining 2-5 outcomes (hot lead, nurture, not a fit), and the AI routes each contact based on their form data, conversation history, and custom fields. Start with lead qualification after a form trigger, define specific criteria in your prompt, and test with real contacts before going live.
🎯 Key Takeaways
✔The AI Decision Maker reads context u2014 conversation history, custom fields, tags u2014 not just field values. Use it when your routing decision requires judgment, not just matching.
✔Write prompts with numbered outcomes and specific criteria (budget ranges, location, urgency signals). Vague prompts produce inconsistent routing.
✔Always include a catch-all branch for contacts with insufficient data u2014 this prevents edge cases from falling into the wrong sequence.
✔Pair Conversation AI (for data collection) with AI Decision Maker (for routing) u2014 they're designed to work together and this combination is where the biggest efficiency gains happen.
✔Test with 10-15 real contacts manually before activating. Include edge cases u2014 incomplete forms, short replies, wrong-fit leads.
✔Review routing accuracy weekly for the first month. Most prompts stabilize after 2-3 refinement iterations within 3-4 weeks.
✔In high-volume lead environments (real estate, agencies, course creators), AI Decision Maker routing typically reduces wasted outbound calls by 40-60% within the first 45 days.
🔍 In-Depth Guide
Writing AI Decision Maker Prompts That Actually Work
The prompt is everything. I've reviewed dozens of GHL workflows where the AI Decision Maker was set up but producing garbage routing u2014 and in 90% of cases, the prompt was vague. Something like 'Is this lead interested?' gives the AI nothing to work with. Every lead is technically 'interested' u2014 they filled out your form.nnWhat works: write the prompt like you're briefing a smart junior analyst who has access to the contact's full file. Define your outcomes with specific criteria. For example: 'Outcome A u2014 contact has specified a budget above AED 500,000, expressed urgency in their messages, and is based in Dubai or Abu Dhabi. Outcome B u2014 contact shows interest but budget or location is unclear. Outcome C u2014 contact is outside the UAE or budget is below AED 200,000.'nnNumber your outcomes. Match those numbers exactly to your workflow branch names u2014 GHL uses the labels to route. Always add a catch-all Outcome D for 'unclear or insufficient data.' Without it, edge cases fall through. I recommend reviewing routing decisions weekly for the first month and tweaking one variable at a time until accuracy hits 85%+.
Using AI Decision Maker for Dubai Real Estate Lead Qualification
Real estate is where I've seen this tool perform best, especially in Dubai's fast-moving off-plan market. A typical high-volume developer or agency is dealing with leads from 10+ countries, three languages, wildly different budgets, and a mix of end-users and investors u2014 all coming in through the same Facebook form.nnHere's what a working setup looks like: Trigger is a Facebook Lead Ad submission. First action is a Conversation AI chat that asks two qualifying questions u2014 budget range and purpose (own use or investment). After that exchange, the AI Decision Maker analyzes the form data plus the chat replies and routes to four branches: hot investor (AED 1M+, investment intent) u2192 immediate call from senior broker; warm buyer (AED 500Ku20131M) u2192 5-day WhatsApp nurture; cold inquiry (below AED 300K or vague) u2192 low-touch email sequence; non-UAE based u2192 region-specific email with international buying guide.nnThis setup alone eliminated about 60% of wasted outbound calls for one client. Their cost per qualified appointment dropped by roughly half within 45 days. The key was pairing Conversation AI for data collection with AI Decision Maker for routing u2014 they're designed to work together.
Common Mistakes to Avoid When Configuring AI Decision Maker
A mistake I see constantly: overloading the context field with every piece of data on the contact. More data is not always better. If you feed the AI 15 custom fields, half of which are empty or irrelevant, the decisions get noisy. Include only the 3-5 data points that actually matter for the decision you're asking it to make.nnSecond mistake: not testing with real edge cases before going live. Every workflow has that weird contact u2014 the one who fills out the form in all caps, replies in a different language, or leaves half the fields blank. Run 10-15 contacts through manually using the Test Workflow feature before turning it on. If the AI routes a blank-field contact to your 'hot lead' branch, your prompt has a gap.nnThird: not iterating. Your first prompt won't be your best prompt. After the first week live, pull the contacts that routed incorrectly and ask yourself u2014 what did my prompt fail to account for? Add a line. Narrow a criteria. I usually lock in a stable prompt by week 3-4. Action you can take today: open one existing workflow, add an AI Decision Maker after your lead form trigger, and write a three-outcome qualification prompt specific to your niche. Run your last 10 leads through it manually and see how it performs.
Most GoHighLevel users I meet are still stuck in the stone age of automation — if this tag, do that. If field equals X, send email Y. That logic works fine for simple sequences. But the moment a lead sends a mixed-signal reply, or a prospect fills out your form with vague answers, the whole thing falls apart. That’s exactly the problem the AI Decision Maker was built to solve.The GHL AI Decision Maker is a workflow action that reads contact data — conversation history, custom fields, tags, pipeline stage — and makes a contextual routing decision the way a trained sales rep would. Not binary logic. Actual judgment. You define the possible outcomes (A, B, or C), write a clear prompt explaining what those outcomes mean, and the AI routes each contact to the right branch automatically. It launched as part of GHL’s 2025 platform update and has quietly become the most powerful feature most agency owners haven’t configured properly yet.I’ve been training agents and agency owners in Dubai on GoHighLevel since 2022, and the number one problem I see is teams spending hours manually triaging leads that should have been sorted automatically. A real estate developer I work with was getting 400+ leads a month from Meta ads. Their team was calling everyone — ready buyers, tire-kickers, wrong country, wrong budget. After we built a two-stage AI Decision Maker workflow, only 20% of those leads went to their sales team, and conversion rate on those calls jumped from 11% to 31%. The AI wasn’t perfect on day one. But after two weeks of prompt refinement, it was better than their junior sales staff.What makes this different from standard conditions is context. A standard GHL condition checks if a field value matches a rule. The AI Decision Maker reads everything you give it and synthesizes a decision. You can literally say: “Based on this contact’s form answers, their budget field, and the last 3 messages in their conversation, decide if they’re ready to buy, need nurturing, or aren’t a fit.” That’s not a filter — that’s a qualification call. And it runs in seconds, at scale, for every contact in your pipeline.
❓ Frequently Asked Questions
The GoHighLevel AI Decision Maker is a workflow action that uses AI to analyze contact data u2014 including conversation history, custom fields, tags, and pipeline stage u2014 and routes leads to different workflow branches based on contextual reasoning. Unlike standard GHL conditions that use binary if/else logic, the AI Decision Maker interprets nuanced scenarios like sentiment, buyer readiness, or lead fit. You write a prompt defining 2-5 possible outcomes, and the AI assigns each contact to the most appropriate branch automatically. It was introduced in GHL's 2025 platform update.
Standard GHL conditions check whether a field value matches a specific rule u2014 for example, 'if tag contains Interested, send email.' They're precise but rigid. The AI Decision Maker reads multiple data points together and makes a holistic judgment call. For instance, it can determine whether a lead who replied 'maybe later' to a sales message is worth a follow-up call or should go into a long-term nurture u2014 something a binary condition can't assess. The AI Decision Maker is best used when the routing decision requires context and interpretation rather than a simple field match.
You can configure between 2 and 5 outcome branches in the AI Decision Maker. The GHL interface creates a separate workflow branch for each outcome you define in your prompt. Most practitioners find 3-4 branches is the sweet spot u2014 enough to segment meaningfully without making the prompt too complex for consistent accuracy. Always include one catch-all branch (e.g., 'Unclear u2014 needs manual review') to handle contacts where the AI has insufficient data to make a confident decision.
Yes. When you configure the Context section of the AI Decision Maker, you can include the last SMS message content, email replies, and conversation history from GHL's Conversation AI chats. This is one of its most powerful features u2014 it means the routing decision can factor in what a lead actually said, not just what they filled in on a form. For example, after a Conversation AI chat session, the AI Decision Maker can assess sentiment from the full exchange and route accordingly u2014 positive replies to a sales team, negative replies to an escalation workflow.
The highest-impact use cases are: (1) lead qualification routing after form submissions or Conversation AI chats u2014 sorting hot, warm, and cold leads automatically; (2) post-appointment follow-up routing based on call notes added to custom fields; (3) re-engagement campaigns where you need to decide whether an old lead is worth a phone call, an email, or should be archived; and (4) sentiment-based routing after SMS or chat conversations. In real estate and high-ticket service businesses especially, replacing manual lead triage with AI Decision Maker routing typically reduces wasted sales calls by 40-60%.
A strong prompt names each outcome clearly (Outcome A, B, C), defines specific criteria for each (budget range, urgency signals, geographic fit), references the exact custom fields you're including in the context, and ends with instructions for what to do when data is ambiguous. Avoid vague prompts like 'Is this lead interested?' u2014 they produce inconsistent routing. Be as specific as you would be if briefing a new sales team member. For example: 'Outcome A: contact has stated a budget above $5,000 and expressed urgency in their last message. Outcome B: contact shows interest but hasn't mentioned budget. Outcome C: contact appears to be a student or not a business owner.' Test with 10 real contacts before going live.
No u2014 they serve different functions and work best together. Conversation AI is a chatbot that talks directly with leads via SMS or web chat in real time. The AI Decision Maker is a backend routing tool that never interacts with the lead u2014 it analyzes data and determines which workflow path the contact should follow. A common and effective setup is: Conversation AI collects qualifying information from the lead, then the AI Decision Maker reads that conversation plus form data and routes the contact to the appropriate follow-up sequence. Think of Conversation AI as the intake rep and AI Decision Maker as the sales manager deciding what happens next.
📘
New Book by Sawan Kumar
The AI-Proof Marketer
Master the 5 skills that keep you indispensable when AI handles everything else.
