Example app to show usage of the "AI skills" app feature in monday.com.
It demonstrates how to build a workflow block to interact with the AI sidekick correctly.
Internal docs are here: https://developer.monday.com/apps/docs/skill-audhs123
When an AI skill runs, it will execute a workflow block under the hood. This way, developers can reuse existing integration functionality in both monday workflows and sidekick.
sequenceDiagram
participant User
participant monday sidekick
participant Skill app feature
participant Workflow block app feature
User ->> monday sidekick: "Give me advice to improve <br/>my email marketing campaigns"
monday sidekick -->> Skill app feature: 1. Interprets user input<br/>2. Chooses a skill to run<br/>3. Pass input data to skill
Skill app feature -->> Workflow block app feature: Runs workflow block action with inputs:<br/> "Mailchimp - get advice"
Workflow block app feature -->> Workflow block app feature: Get advice from Mailchimp API
Workflow block app feature -->> monday sidekick: Returns action state and output data
monday sidekick -->> monday sidekick: Interprets block output and<br/>prepares response to user
monday sidekick ->> User: Here's what you do well in your emails, <br/>and this is what you can improve...
The repo contains Node.js and Python examples. It calls the Github API and the Mailchimp API.
| Skill | Industry | Demonstrates | Supported actions | Languages | Ways to extend the app |
|---|---|---|---|---|---|
| Github repo management skill | Software or IT permissions management | Action block, store authentication keys, make external API call | Get github repos, get github repo collaborators, add collaborator to repo | NodeJS, Python | Support more admin actions in Github, or update permissions in more apps (eg Okta) |
| Mailchimp email insights skill | Marketing and customer success | Action block, make external API call | Get mailchimp campaigns, get campaign metadata, get campaign insights and advice | NodeJS | Get insights from mailing list subscribers in Mailchimp, integrate with data enrichment tools, text analysis |
Open the Node.js or Python folder and follow instructions in the README.md file.