Unit Introduction
Welcome to the third and final unit of the MCP course.
You have just learned about Make MCP client. Now, let’s look at the MCP server.
In this unit, you will learn: what you can do with the Make MCP server
when to use it and what benefits it offers
how to use the Make MCP server
What is it?
Make provides an MCP server.
The Make MCP server allows external AI systems, such as AI agents or AI chats, to run your scenarios and manage the contents of your Make account.
This means that Make has implemented MCP on its side to enable AI models to connect to it and access specific Make functions. AI systems can run your Make scenarios as tools and they can also manage your teams, organizations, and scenarios. AI models can directly trigger and manage your Make scenarios using natural language. Instead of manually running scenarios, using the Make interface or the Make API, you can simply ask an AI system to do it for you.
You can give the following instruction to your AI system: when a new lead comes in with revenue over $100K, run my enterprise onboarding scenario and notify the sales team. The AI understands the task, finds the relevant scenarios, and executes them.
🗣️ These are the benefits of using Make as an MCP server:
Conversational automation Tell an AI to perform an action, and the AI runs the corresponding scenario. This eliminates the need to click through the Make interface, remember scenario names, or navigate dashboards.
🎸 Intelligent orchestration AI can decide which scenario to run or run multiple scenarios based on your requests.
🚀 Faster workflow management 🚀 The MCP host makes sure each MCP client is properly connected to the MCP servers it
needs. It keeps track of which clients are active and ensures they can communicate
properly with MCP servers.
Tools
Make as an MCP server offers three kinds of capabilities:
1 Run active and on-demand scenarios
2 View and modify scenarios and their related entities
3 View and modify teams and organizations
Let’s see each one in detail.
1 : Runactiveandon-demandscenarios
Run active and on-demand scenarios
Your AI model can use every Make scenario that you set to active and on-
demand as a tool.
When the AI identifies a task that matches one of your scenarios, it can trigger it without manual intervention. This means your Make scenarios become tools that the AI can use automatically whenever it needs to complete a specific task.
If you have a scenario that sends email reports, the AI can run it when performing a task that requires sharing a report.
View and modify scenarios and their related entities
Make offers the AI systems tools to mange scenario and related entities. Your
AI model can see and edit your Make scenarios along with connections,
webhooks, and data stores.
This gives the AI the capability to make changes to your scenario settings. When the AI needs to update a scenario, it can access and adjust the settings directly. This means you don’t have to manually dig through configurations yourself. You ask the AI to update a data store record and the AI can modify the record directly using the Update data store record detail tool that Make provides.
3: View and modify teams and organizations
View and modify teams and organizations
Make offers the AI systems tools to modify teams and organizations. Your AI
model can see and manage all assets within your Make organizations, plus
access details about the organizations themselves. This gives the AI the capability to understand your organization structure and make changes automatically, so you don’t have to navigate through account settings manually. Whether it needs to verify settings or make changes to your organizations and teams, the AI can use the relevant tools to handle tasks like renaming teams, updating timezone and region preferences, creating new teams, and more.
You can ask the AI to update timezone settings in the chat, and it can handle that directly without you needing to update it manually.
You might have noticed that most of the actions that the Make MCP server lets you perform are the same actions that the Make API provides. MCP, however, makes these capabilities accessible through natural language conversation rather than requiring you to write code or make direct API calls. You don’t need to know the specific API endpoints and request formats. You can simply describe what you want to accomplish in plain language and let the AI handle the technical implementation. This makes everything easier and quicker, even for users without a technical background. When connecting an AI system to the Make MCP server, you choose which permissions (scopes) to grant, which determines what tools the AI can access. This
gives you a level of control of what the AI can access and do, similar to what you have with APIs.
For example, you might grant permission to run scenarios but not to modify them, or allow viewing team members but not changing their roles. Note that all Make plans can use scenarios as AI tools, while management tools (for teams, organizations, and scenario editing) are only available on paid plans.
When to use it
Use Make as an MCP server when you want AI systems to interact with your workflows and run or manage them easily.
🧱 By using Make as your MCP server, you can:
Create custom tools Group several steps (like sending messages or updating data) into one simple tool.
🕵️♀️ Hide sensitive data You can choose which data to share with the AI or hide sensitive info.
✅ Add approval steps and logs All actions can be tracked, and you can require manual approvals for sensitive steps.
🔑 Give AI agents only the access they need Y d id tl hi h ti th AI f d htdtit 🔑 You decide exactly which actions the AI can perform and what data it can access.
☎️ Simplify calls to third-party applications You only need to provide the fields relevant to your use case.
When can you use the Make MCP server in real life?
Let’s have a look at three real-life examples.
1 🗣️ Voice-Activated help desk AI agent 2 💼 AI sales assistant 3 👁️ Security AI agent
1 : Voice-ActivatedhelpdeskAIagent
🗣️ Voice-Activated help desk AI agent 🎯 Objective Build a voice-activated AI agent for IT support where users can call or send voice messages to request help, with the AI handling ticket creation, status checks, and team notifications.
🏗️ How to do it Create three Make scenarios and expose them as MCP tools:
Create a Jira ticket and send a Slack notification Search Jira tickets by keywords Get specific ticket details.
Connect your voice AI platform (VoiceFlow, VAPI, etc.) to Make’s MCP server so the AI agent can call these scenarios during conversations. 🎬 See it in action A client calls: My laptop won’t connect to Wi-Fi.
The AI calls the create ticket tool to generate Jira ticket #1234 and notify the IT team via Slack.
Later, a support agent asks: What’s the status of #1234? The AI calls the get ticket details tool to retrieve the current information instantly.
💜 Advantage of using Make Make groups multiple steps into one simple tool: the AI uses a single tool to create a Jira ticket and notify Slack. Make hides Jira’s complexity by only exposing relevant fields (title, description, priority) instead of requiring the AI to handle dozens of technical parameters.
2: AI sales assistant
💼 AI sales assistant 🎯 Objective Enable a sales engineer to use their AI chat (Claude, ChatGPT) to perform daily CRM tasks like checking opportunities, adding comments to deals, retrieving contacts, and generating reports through natural conversation.
🏗️ How to do it Create four on-demand Make scenarios as MCP tools:
List sales opportunities from Salesforce Add comments to opportunities Create Google Docs List contacts in opportunities.
Connect Make’s MCP server to your AI chat application to allow it to call these scenarios.
🎬 See it in action A sales engineer asks their AI chat: What deals are closing this quarter?
The AI calls the list opportunities tool to retrieve filtered Salesforce data and displays the results.
They then say: Add a note to the Acme Corp deal that we discussed pricing today, and the AI calls the add comments tool to update the opportunity instantly.
💜 AdtfiMk 💜 Advantage of using Make Make allows you to choose exactly which CRM data the AI can access, you can hide sensitive fields, filter by region or deal size, and protect confidential client information. With Make you can simplify Salesforce’s complex parameters into easy fields the AI understands. Make allows you to track all AI actions with built-in logging, and optionally require approval for high-value deal updates.
3: Security AI agent
👁️ Security AI agent 🎯 Objective Build an AI agent to automate sensitive IT security and compliance tasks like managing employee accounts, performing password resets, auditing role conflicts, and onboarding new hires while maintaining strict security and approval controls.
🏗️ How to do it Create Make scenarios as controlled MCP tools:
List active employees from Workday List identity provider employees from Entra ID with their roles and status Change employee roles Send welcome emails.
You can implement logging, approval workflows, and data filtering within each scenario to track all actions and require authorization before executing sensitive changes.
🎬 See it in action Security admin asks: Check if any employees have conflicting admin roles.
The AI calls the list Entra ID tool to retrieve employee roles, analyzes them, and flags potential conflicts.
The admin then asks: Onboard John Smith with standard employee access and the AI calls the change roles tool, which triggers an approval workflow. After approval, it updates permissions and calls the welcome email tool automatically.
💜 Advantage of using Make Make gives you complete security control over sensitive operations. You can add mandatory approval workflows before any role changes, filter out confidential employee data, and maintain detailed logs of every AI action for compliance. With Make you can combine complex security operations (verify permissions, update roles, notify employee) into single, trackable tools instead of multiple API calls. single, trackable tools instead of multiple API calls. Make protects your organization by letting you control exactly what the AI can access and requiring human approval for sensitive actions, something direct MCP tools can’t guarantee.
How to use it
To use Make as an MCP server, you first need to connect it to your AI system.
You can use an AI agent or any AI system that implements MCP, like Cursor, Claude, or ChatGPT. There are two ways to connect according to the authentication method you choose:
1 OAuth
2 Token
Let’s see each one in detail. 1 : OA u t h
OAuth
It is the recommended method due to its simple and secure setup.
Check out the instructions to connect your AI system to the Make MCP server using OAuth
TAKE ME THERE! Click the image to enlarge
During the OAuth connection process, you’ll see a consent screen where you
select which organization to connect and which scopes to grant. These scopes
determine what tools the AI system can access and what it’s allowed to do. This
gives you control over what data and actions are available to your AI system. When connecting via OAuth, AI access is restricted to the single organization you select during setup. By default, it can use all on-demand and active scenarios within that organization.
To limit access further, a Teams plan lets you restrict access to only scenarios within specific teams. You can accomplish this by setting up separate user accounts with team-specific access. Connecting through these accounts ensures that their permission restrictions are applied accordingly.
2 : To ke n
Token
Make allows you to use tokens as well to authenticate your connection. You just have to generate a token from your Make account settings and use it directly in your MCP configuration. Check out the instructions to connect your AI system to the Make MCP server using tokens
TAKE ME THERE! When selecting the scopes, you need to include mcp:use to make your on- demand and active scenarios available as tools. You can also add additional scopes to allow your AI to perform other actions on
your organization or scenarios.
Note that by default, the mcp:use scope allows AI systems to access all active and on-demand scenarios across all of your Make organizations. You can restrict access at the organization, team, or scenario level by following these instructions.
Choose the method that works best for your security needs, then set up the connection in your AI system.
The documentation includes instructions for connecting Make as an MCP server to Cursor, Claude, or ChatGPT, though you can connect it to any AI system that supports MCP.
See it in action
Let’s look at two examples of using Make as an MCP server. You’ll see how to call scenarios as tools and how to use AI to manage your Make organizations.
Scenarios as tools
Query and update a CRM using natural language.
A sales manager uses the following software to manage their work:
Airtable
Table 1 to store lead information. It contains their contact details, their status (e.g., qualified lead, customer, etc.), and the name of the employee responsible for each one. Table 2 with the information of the employees.
Asana
Store all the tasks related to each lead/customer.
They want to manage both software from their AI chat using natural language. To do so, they build three Make scenarios.
Work through each stage before you continue.
Scenario 1 - List contacts
Returns the list of the contacts in Airtable
Scenario 2 - Contact information
Returns information for a specific contact, including details about the assigned employee and related tasks. Scenario 3 - Update contact type
Change the type of contact in Airtable (e.g. from qualified lead to customer).
Let’s see it in action using Claude.
The first step is to create a connection between Claude and Make by following the instructions in the documentation. When setting up the connection, you can define whether you want the AI system to ask for approval when calling specific tools, and select which tools the AI can access.
Once everything is set up, you can run your tools directly from Claude.
Let’s have a look.
Work through each stage before you continue. Step 1
Get contacts
You can ask the AI system to list all contacts of a specific type. The AI calls Scenario 1 and returns a list of all the contacts. Step 2
Get info
You can ask the AI system to get more information about a specific contact. The AI calls Scenario 2 and returns all the information about the contact you requested. You can also see that the AI suggests that the contact should change status from qualified lead to customer. Step 3
Change contact type
You can ask the AI system to change the type of a specific contact.
The AI calls Scenario 3 and performs the change.
By using Make as an MCP server, you can run your scenarios as tools from your AI chat. The AI automatically decides which scenario tool to use based on the request.
🗣️ Here are the advantages:
Riitll 🗣️ Run your scenarios using natural language.
🧱 Combine multiple apps into a single tool.
🔑 Control exactly what the AI can access.
Manage your organization
Manage your Make organization from your AI chat. You can use Claude to manage your Make organization and scenarios. Simply
create a connection as described above and choose which organization the AI can access. Let’s see what you can do.
Work through each stage before you continue. Step 1
List teams and list scenarios
You can as the AI system to list all the teams in the selected organization and all its scenarios. Step 2
Explain the function of a scenario
You can ask the AI system to explain what a specific scenario does. The AI will retrieve the scenario’s blueprint, analyze what it does, and provide an explanation. Step 3
Organize scenarios into folders
You can ask the AI system to categorize your scenarios by what they do, create corresponding folders, and move each scenario into the appropriate one.
By using Make as an MCP server, you can manage your Make account by using natural language and without needing to log in to Make.
🔑 Here are the advantages:
Choose which organization the AI can access and decide what it can do.
✅ ✅ Add approval steps and logs if you’re doing sensitive operations.
Excellent! You’ve seen what’s possible with Make as an MCP server.
Continue to complete this unit
Wrap up
Make can act as an MCP server, letting AI systems manage your workflows through conversation. You can use your AI to run scenarios as automated tools, manage and organize your workflows, and control your organizations and teams, all using natural language and from your AI chat.
Using Make as an MCP server changes how you manage automations. You can control everything with natural language: from running to managing your scenarios. The advantage of using Make is that you can build custom tools that contain calls to different third-party applications. You can also stay fully in charge of security by deciding what the AI can access, hiding sensitive data, adding approval steps when needed, and keeping detailed activity logs.
To use Make as an MCP server, start by connecting your AI system using either OAuth or token authentication.
During setup, you’ll specify which Make organization your AI should access and what actions it’s allowed to perform, giving you complete control on what the AI can do. By default, AI gets access to all your active and on-demand scenarios, but you can restrict this to specific organizations, teams, or individual scenarios.
Once configured, you’re ready to start running scenarios and managing your Make setup directly from your AI chat. This works with Claude, Cursor, ChatGPT, and any AI system that supports MCP. Great job! You now know how to use Make as an MCP server.
That’s it for this course.
You can continue learning about AI agents in Make or start applying what you’ve learned in your own projects.
Mark this task complete to finish this course.