How do you build an MQL scoring workflow in Make.com?

Quick Answer: You can build an MQL scoring workflow in Make.com by connecting your CRM and marketing tools, pulling in behavioural signals (page visits, email clicks, demo requests), running them through an AI logic step to calculate a weighted score, and automatically tagging contacts when they cross their MQL threshold. The whole workflow runs without manual intervention.

Building a lead scoring model that actually works requires more than assigning arbitrary point values in a spreadsheet. Most teams set up a basic scoring system, watch it drift out of sync with reality, and end up with sales reps chasing leads that were never ready. This tutorial shows you how to build a behaviour-driven MQL scoring workflow in Make.com that scores contacts in real time, applies AI logic to weight signals intelligently, and routes qualified leads to your CRM automatically.

By the end, you will have a working Make.com scenario that tags and scores MQLs without touching a single lead manually.

What Is an MQL Scoring Workflow?

An MQL scoring workflow is an automated process that assigns numerical scores to contacts based on their behaviour and firmographic fit, then classifies them as Marketing Qualified Leads when they cross a defined threshold.

The workflow monitors signals like email engagement, page visits, content downloads, and demo requests. Each signal carries a weighted score. When a contact's cumulative score passes your MQL threshold, the system tags them, updates your CRM, and (optionally) alerts your sales team.

Make.com is well-suited for this because it connects to virtually every marketing and CRM tool, supports conditional logic natively, and allows you to drop an AI module into the scoring step to handle nuanced weighting that static point systems miss. If you are comparing this with broader channel orchestration or campaign execution support, SaaS Hackers also curates a list of the best B2B SaaS digital marketing agencies.

Why Static Lead Scoring Breaks Down

Static lead scoring assigns fixed points to actions: 10 points for opening an email, 20 for visiting the pricing page, 5 for downloading a guide. The problem is that not all signals carry equal weight across different contact profiles.

A contact who visits your pricing page twice in one week is a stronger signal than one who opened three newsletters. A contact from a 500-person SaaS company hitting your pricing page means something different than a freelancer doing the same thing.

AI logic in Make.com lets you pass contextual data into a scoring prompt and get back a weighted score that accounts for signal combination, recency, and firmographic fit simultaneously. That is the key difference between this approach and a basic point-tally system. Teams that want to combine lead scoring with targeted account selection often pair this kind of workflow with support from B2B SaaS ABM agencies.

What You Need Before You Start

Before building the workflow, get these in place:

• A Make.com account (Core plan or above for multi-step scenarios)
• A CRM connected to Make.com (HubSpot, Attio, Salesforce, or Pipedrive all work)
• An event source where behaviour signals arrive (your CRM, a form tool, or a webhook from your website)
• An OpenAI API key (for the AI scoring step)
• A defined MQL threshold agreed with your sales team (e.g., 70 points out of 100)
• A list of scoring signals with rough weights (you will refine these with AI logic)

If you have not aligned with sales on your MQL threshold yet, do that first. The threshold is the single most important variable in the model. Sojourn Solutions describes it well: the MQL threshold is the score at which a lead gets classified as marketing-qualified and routed to sales, and setting it incorrectly is the most common reason scoring systems fail.

How to Build the MQL Scoring Workflow in Make.com

Step 1: Set Up Your Trigger Module

Your workflow needs a trigger that fires whenever a relevant contact event occurs.

In Make.com, create a new scenario and add your trigger module:

• HubSpot: Use "Watch Contacts" or "Watch CRM Events" to trigger on property changes or form submissions
• Webhook: If your website fires events directly, use Make's "Custom Webhook" module as the trigger
• Typeform / Tally: Use their native Make.com modules to trigger on form completions

Set the trigger to fire on the events most relevant to your funnel: demo requests, pricing page visits, email link clicks, content downloads, or free trial sign-ups.

Map the following fields from your trigger payload:

• Contact email or ID
• Event type (e.g., "pricing_page_visit", "demo_request")
• Timestamp
• Company name and size (if available)
• Any existing score already stored in your CRM

Step 2: Retrieve the Contact's Current Data

Add a second module to pull the full contact record from your CRM. This gives you access to:

• Their current lead score (if one exists)
• Firmographic data (company size, industry, job title)
• Previous activity history
• Current lifecycle stage

In HubSpot, use the "Get a Contact" module with the email or contact ID from Step 1. In Attio, use "Get Record". In Pipedrive, use "Get Person".

You need this data so the AI step in Step 4 has full context, not just the single event that fired the trigger.

Step 3: Aggregate Behavioural Signals with a Data Store

Make.com's Data Store module acts as a lightweight database. Use it to accumulate behavioural events per contact before scoring.

Set up a Data Store with the following fields:

Each time the workflow runs, use "Search Records" to find the contact's existing entry, then use "Update a Record" (or "Add a Record" if none exists) to log the new event.

This aggregation step means your AI scoring module receives a full behavioural summary, not just a single event signal. If content engagement is one of your strongest MQL inputs, it can also help to review the best B2B SaaS content marketing agencies for ideas on what signals to weight more heavily.

Step 4: Score the Lead with an AI Logic Step

This is where Make.com's workflow goes beyond static scoring. Add an OpenAI "Create a Completion" module (or the "Create a Chat Completion" module for GPT-4o).

Write a system prompt that instructs the model to act as a lead scoring engine. Here is a working prompt structure:

You are a B2B SaaS lead scoring engine. Score the following contact 
out of 100 based on their behaviour and firmographic data.

Scoring criteria:
- Demo request: 35 points
- Pricing page visit (per visit, max 3): 15 points each
- Email link click (per click, max 5): 5 points each
- Content download: 10 points
- Free trial sign-up: 40 points
- Company size 50-500 employees: +10 bonus
- Job title contains "Head", "VP", "Director", or "C-": +10 bonus
- Events older than 30 days: reduce score by 20%

Contact data:
- Event type: {{event_type}}
- Event count: {{event_count}}
- Last event date: {{last_event_date}}
- Company size: {{company_size}}
- Job title: {{job_title}}
- Existing score: {{current_score}}

Return only a JSON object in this format:
{"score": [number], "mql": [true/false], "reason": "[brief explanation]"}

Map the variables from your Data Store output into the prompt using Make.com's variable mapping panel.

Set the model to GPT-4o mini for cost efficiency on high-volume workflows, or GPT-4o if you need more nuanced reasoning on complex scoring rules.

Step 5: Parse the AI Response

Add a JSON "Parse JSON" module after the OpenAI step. Point it at the choices[0].message.content output from the AI module.

This extracts three values you will use in the next steps:

• score (the numerical score)
• mql (true or false)
• reason (a plain-English explanation of the score)

If the AI occasionally returns malformed JSON, add a Tools "Set Variable" module with a fallback value of {"score": 0, "mql": false, "reason": "parse error"} and route errors there before continuing.

Step 6: Add a Router for MQL vs Non-MQL Paths

Add a Router module after the JSON parser. Set up two paths:

Path 1: MQL = true

• Filter condition: mql equals true

Path 2: MQL = false

• Filter condition: mql equals false

This splits the workflow so qualified leads get different treatment from leads still in nurture.

Step 7: Update the CRM and Tag MQLs

On the MQL path:

1. Add a "Update Contact" module in your CRM. Set the lead score property to the score value from the AI output.
2. Set the lifecycle stage to "Marketing Qualified Lead" (or your CRM's equivalent).
3. Add a tag or label: "MQL - Auto-scored".
4. Optionally, add a Slack or email notification module to alert the relevant sales rep.

On the non-MQL path:

1. Update the contact's score field with the new score value.
2. Keep the lifecycle stage unchanged.
3. Optionally, enrol them in a nurture sequence if their score crossed an intermediate threshold (e.g., 40-69 points).

Step 8: Store the Score History

After both paths, add a final Data Store "Update a Record" step to write the new score back to your aggregation store. This ensures the next time the workflow fires for this contact, the AI step starts from the correct baseline score rather than zero.

How to Set Your MQL Threshold Correctly

The MQL threshold is the score at which a contact gets classified as sales-ready. Set it too low and sales gets flooded with unqualified leads. Set it too high and your pipeline dries up.

Use this process to find the right number:

1. Pull 3-6 months of closed-won deals from your CRM
2. Identify the behavioural signals those contacts showed before they became customers
3. Run those signals through your scoring model and note what score they would have received
4. Find the score that captures 80% of your closed-won contacts without pulling in more than 30% of contacts who never converted

Start with a threshold of 65-75 for most B2B SaaS models. Review it every 90 days against actual conversion data. If your scoring model is closely tied to inbound motion, the top B2B SaaS inbound marketing agencies list may be useful for benchmarking how other teams structure qualification.

Testing Your Make.com MQL Scoring Workflow

Before going live, run the scenario in test mode with known contacts:

• Use a contact who has already converted as a customer. Their score should exceed your MQL threshold.
• Use a cold contact with no activity. Their score should sit below 20.
• Use a contact with mixed signals (a few email opens, no high-intent actions). Their score should land in the mid-range.

Check the AI output in the scenario execution log. If the scores feel off, adjust the prompt weights in Step 4 and re-run.

Make.com's execution history shows every module input and output, so you can trace exactly where a score came from and why.

Common Mistakes to Avoid

Skipping the data aggregation step. Scoring on a single event in isolation produces wildly inaccurate results. Always aggregate signals before scoring.

Not aligning the threshold with sales. A technically perfect scoring model fails if sales does not trust the output. Show them the scoring logic and get explicit agreement on the threshold before launch.

Ignoring score decay. A contact who visited your pricing page six months ago is not the same signal as one who visited yesterday. The prompt in Step 4 includes a 30-day decay rule. Adjust the timeframe to match your sales cycle length.

Over-engineering the AI prompt. More rules in the prompt do not always produce better scores. Start with 5-7 signal types, run the model for 30 days, then add complexity based on what the data shows.

FAQs

What is an MQL scoring workflow in Make.com? An MQL scoring workflow in Make.com is an automated scenario that collects behavioural signals from your CRM and marketing tools, passes them through a scoring logic step (including an AI module for weighted scoring), and automatically tags contacts as Marketing Qualified Leads when they cross a defined score threshold. It removes manual lead qualification from the process entirely.

How do I set the right MQL score threshold in Make.com? Pull your last 3-6 months of closed-won deals, identify what behavioural signals those contacts showed before converting, simulate their scores through your model, and find the score that captures at least 80% of them. For most B2B SaaS businesses, the MQL threshold falls between 65 and 75 out of 100. Review it every 90 days against actual pipeline conversion data.

Can Make.com replace a dedicated lead scoring tool like HubSpot's native scoring? Make.com does not replace your CRM's native scoring for simple models, but it adds meaningful capability when you need AI-weighted scoring, cross-tool signal aggregation, or custom routing logic that native tools do not support. If your scoring model needs to pull signals from more than one platform or apply contextual weighting, Make.com gives you more control than most native implementations. For teams extending native CRM workflows, SaaS Hackers also has a directory of B2B SaaS HubSpot agencies.

What CRMs work with this Make.com MQL scoring workflow? HubSpot, Salesforce, Pipedrive, Attio, and Zoho CRM all have native Make.com modules that support contact retrieval and update operations. The workflow structure in this guide works with any of them. You swap the CRM-specific modules in Steps 2 and 7 while keeping the AI logic and routing steps identical.

How much does it cost to run this workflow in Make.com? The main variable cost is OpenAI API usage. Running GPT-4o mini for scoring completions costs approximately $0.15 per 1 million input tokens. For a workflow processing 1,000 contacts per month with a prompt of roughly 300 tokens each, the AI cost is under $0.05 per month. Make.com operation costs depend on your plan, with each module execution counting as one operation.