What?

Zoho has released new resources to help AI coding assistants build solutions for Zoho CRM. These cover tasks such as retrieving records, automating work and creating custom features.

There are two main parts:

  • SKILL.md is an instruction file that helps an AI assistant choose the right Zoho CRM tools and use the supporting documentation. The .md extension means Markdown: a plain-text format with simple headings and lists.
  • OAS stands for OpenAPI Specification. An API lets programs communicate with each other. These specification files describe the requests Zoho CRM accepts, the information you need to send, and the responses you can expect.

Think of the skill as a guide for deciding how to approach a task, and the OAS files as the detailed reference for making API requests.

Zoho has also reorganised the API repository. The specifications now use YAML, a readable format for storing structured information. Each API area has its own folder with an overview, details of individual operations, and request and response examples. This makes the relevant information easier to find.

What?

Zoho has released new resources to help AI coding assistants build solutions for Zoho CRM. These cover tasks such as retrieving records, automating work and creating custom features.

There are two main parts:

  • SKILL.md is an instruction file that helps an AI assistant choose the right Zoho CRM tools and use the supporting documentation. The .md extension means Markdown: a plain-text format with simple headings and lists.
  • OAS stands for OpenAPI Specification. An API lets programs communicate with each other. These specification files describe the requests Zoho CRM accepts, the information you need to send, and the responses you can expect.

Think of the skill as a guide for deciding how to approach a task, and the OAS files as the detailed reference for making API requests.

Zoho has also reorganised the API repository. The specifications now use YAML, a readable format for storing structured information. Each API area has its own folder with an overview, details of individual operations, and request and response examples. This makes the relevant information easier to find.

Why?

AI can produce code that looks convincing but contains mistakes. It might use the wrong field name, misunderstand a Zoho feature or prepare an incorrect API request.

These resources give the assistant official instructions and examples to follow. The aim is to produce a better starting point and reduce the time spent correcting basic errors.

For example, you might ask the assistant to find open deals that are due to close this month. It needs to choose a suitable method, use the correct fields and request the necessary permissions.

Your own CRM setup still matters. Custom fields, different layouts and user permissions can affect the result, so generated code needs checking and testing.

How?

  1. Use an AI coding assistant that supports agent skills. Follow the installation instructions in the Zoho CRM skills repository. With Node.js installed, the basic terminal command is: 
    npx skills add zoho/crm-skills
  2. Restart your assistant after installation and check that the zoho-crm skill is available. Keep the skill and its supporting files together.
  3. Describe one clear task. For example: "Show the names and amounts of open deals due to close this month." Include any relevant module names, field API names and conditions.
  4. Ask the assistant to explain its suggested approach. Review the code or API request, including the permissions it needs. The Zoho CRM API repository provides the specifications and examples for checking API details.
  5. Test the solution in a sandbox: a separate CRM environment used for testing. Use sample records and check that the result matches your request.
  6. If something goes wrong, give the assistant the error message and explain what you expected. Check the revised solution before using it with live records.

For automation functions running on Zoho's servers, the skill currently supports Deluge, Zoho's scripting language. Keep the installed skill updated as Zoho improves its tools and documentation.

Why?

AI can produce code that looks convincing but contains mistakes. It might use the wrong field name, misunderstand a Zoho feature or prepare an incorrect API request.

These resources give the assistant official instructions and examples to follow. The aim is to produce a better starting point and reduce the time spent correcting basic errors.

For example, you might ask the assistant to find open deals that are due to close this month. It needs to choose a suitable method, use the correct fields and request the necessary permissions.

Your own CRM setup still matters. Custom fields, different layouts and user permissions can affect the result, so generated code needs checking and testing.

How?

  1. Use an AI coding assistant that supports agent skills. Follow the installation instructions in the Zoho CRM skills repository. With Node.js installed, the basic terminal command is: 
    npx skills add zoho/crm-skills
  2. Restart your assistant after installation and check that the zoho-crm skill is available. Keep the skill and its supporting files together.
  3. Describe one clear task. For example: "Show the names and amounts of open deals due to close this month." Include any relevant module names, field API names and conditions.
  4. Ask the assistant to explain its suggested approach. Review the code or API request, including the permissions it needs. The Zoho CRM API repository provides the specifications and examples for checking API details.
  5. Test the solution in a sandbox: a separate CRM environment used for testing. Use sample records and check that the result matches your request.
  6. If something goes wrong, give the assistant the error message and explain what you expected. Check the revised solution before using it with live records.

For automation functions running on Zoho's servers, the skill currently supports Deluge, Zoho's scripting language. Keep the installed skill updated as Zoho improves its tools and documentation.