Kaizen 266 - Working with Zoho CRM Using Zoho MCP - Part II
In Part I, we explored how to connect Zoho MCP to an MCP client and use it to work with Zoho CRM records and metadata. We also saw how the client can identify the tools it needs and combine multiple operations based on a natural-language instruction.
In this part, we will take that a step further. We will look at how MCP can be used to customize CRM, configure automation, and even work across multiple Zoho applications.
Customizing Zoho CRM
MCP can also be used to work with the configuration of your CRM and make changes to suit your business requirements.
In this example, we describe a Service Requests module in plain language, including its fields, its relationships with Accounts and Contacts, and who can access it. We then create a global picklist and link it to a field in the module, and finish by creating, fetching and deleting a test record.
Here is a snapshot of the module that we just created:
Points to note:
- A requirement can replace a series of setup steps. The client checks the profiles, then creates the module, its fields and the lookups to Accounts and Contacts from a single description.
- Each step builds on the last. The module, the global picklist, the field that uses it and the test record all come from one conversation, and the client uses what it has just created.
- Test what was built, then clean up. Creating and fetching a test record is a quick way to check the module works. The client asks for permission before creating and before deleting, so you stay in control of changes to your CRM.
Configuring CRM Automation
We can also use MCP to work with automation in Zoho CRM.
In this example, we check which workflow rules exist for Contacts, then create a workflow that adds a follow-up task for the Contact Owner when a new Contact has no email or phone. We then update the same workflow to send an email notification to the owner as well.
- Check before you build. The client first looks at the existing rules and at whether Email and Phone are mandatory, and uses the answers to design the workflow.
- Supporting pieces are created for you. The task, and later the email template and email notification, are created first and then attached to the rule.
- The client validates the result. It spots a due date set in the past and corrects it, and when one call returns an error, it checks the rule's current state before carrying on.
- Updating is just describing the change. Adding the email notification to the existing workflow took one follow-up request.
- Verify in the CRM. For automation, open the rule in your CRM and test it with a sample record before relying on it.
This demonstrates how MCP can be used not only to retrieve existing automation but also to create and modify workflows based on changing requirements. Multiple CRM operations can be combined to build an automation that involves conditions, tasks, email notifications, and other actions.
Working Across Zoho CRM and Zoho Books
The use of MCP is not limited to working within a single application. By configuring tools from multiple Zoho applications on the same MCP server, you can ask an AI agent to handle workflows that involve data from different services.
In this example, we work with Zoho CRM and Zoho Books to automate part of the invoicing process.
Before proceeding, ensure that the Zoho CRM and Zoho Books integration is enabled and that the relevant products and customer records are available in Zoho Books. You should also configure the required CRM and Books tools on the MCP server. Review the invoice details before sending them to customers.
We ask the client to check all Deals marked Closed Won during the current week and see whether each Deal has a corresponding invoice in Zoho Books. If an invoice does not exist, it should create one using the products associated with the Deal, link it to the correct Deal, and leave it in draft mode without sending it.
The client retrieves the Deals from CRM, checks Zoho Books for existing invoices, and fetches the products associated with each Deal through its Products related list. It then matches those products with the corresponding items in Books and creates the invoices.
- One request can involve multiple applications. The client retrieves Deals and their associated products from CRM, checks for existing invoices in Books, and creates invoices where needed. All of this is handled through a single natural-language instruction.
- Records are linked across applications. When creating an invoice, the client passes the corresponding CRM Deal ID so that the invoice is associated with the correct Deal in Zoho Books.
- Check before creating. The client checks for existing invoices before creating new ones, helping avoid duplicate invoices.
- Review before sending. The invoices are initially created as drafts. After reviewing the details, we ask the client to send one selected invoice and verify its status in Zoho Books. The remaining invoices stay in draft mode.
This example demonstrates how MCP can bring together operations from different applications to complete a business workflow. Instead of manually checking Deals in CRM, looking up invoices in Books, and creating the missing ones, you can describe the expected outcome and let the client coordinate the required operations.
Add the Tools you need
The examples above cover only a few of the operations that can be performed through Zoho MCP. The key is that you do not need to configure every available tool. You can add the tools required for your particular use case and connect your MCP server to your preferred MCP client.
For example, if you primarily need to work with CRM records, you can configure the relevant CRM record tools. If you are working on CRM administration or customization, you can add the metadata and configuration tools you need. As we saw in the first example, if the client tells you a tool is missing, you can add it to the server and carry on.
And if your workflow extends beyond CRM, you can configure tools from other Zoho applications or supported third-party services as well. A workflow that would normally require you to work across multiple applications can potentially be handled through a single natural-language instruction, provided the required services and tools are available.
What's Next?
In this two-part series, we have explored how an AI agent can use the Zoho CRM Skill and Zoho MCP to work with CRM. We started with individual CRM operations and then extended the examples to include customization, automation, and workflows involving multiple applications.
The possibilities go beyond the examples covered here. In the coming weeks, we can explore more practical workflows and see how the Zoho CRM Skill and Zoho MCP can help address different real-world business requirements.
If you have any questions, or if there is something specific you would like us to cover in an upcoming Kaizen, leave a comment below or email us at support@zohocrm.com. We would love to hear from you.
Previous Kaizen: Working with Zoho CRM using Zoho MCP - Part I | Kaizen Collection: Home