← All projectsAI Agent · Workflow Automation

AI Agent Jira Bot

A Google Chat agent that runs your sprint board

An AI agent that lives in your team's Google Chat. A Google Chat bot app forwards each request to a backend AI router built on Agent ADK — with MCP tools, workflows, automations, and crons — which reads and writes across Jira, GitLab, Notion, Confluence, Crashlytics, Drive, and Calendar, then replies right back in the chat.

Agent ADK
Google Chat
MCP Tools
Jira
GitLab
Notion
Confluence
Crashlytics
Drive
Calendar

The problem

Engineering managers lose hours a week to sprint-board housekeeping: checking Jira status, cross-referencing GitLab activity, chasing updates across Notion and Confluence, and assembling reports from half a dozen tools. The context is scattered across systems that do not talk to each other, and the person stitching it together is usually the most expensive person on the team.

What we built

The AI Agent Jira Bot lives in your team's Google Chat. Mention it in a channel and it answers in plain language, reading and writing across your whole delivery stack — so running the sprint board becomes a conversation instead of a tab-switching marathon.

Google Chat bot app as the one workspace surface — ask in plain language
Backend AI router on Agent ADK: MCP tools, workflows, automations, and crons
Reads and writes across Jira, GitLab, Notion, Confluence, Crashlytics, Drive, and Calendar
Replies, charts, and sprint reports delivered back into Google Chat

How it is built

A Google Chat bot app forwards each request to a backend AI router built on Google's Agent Development Kit. The router uses MCP tools to reach each system, plus workflows, automations, and scheduled crons for recurring jobs. It reads and writes across Jira, GitLab, Notion, Confluence, Crashlytics, Drive, and Calendar, then replies — with text, charts, or a sprint report — straight back into the chat.

Agent ADK backend AI router
MCP tools for each connected system
Workflows, automations, and scheduled crons
Integrations: Jira, GitLab, Notion, Confluence, Crashlytics, Drive, Calendar

The result

The bot turns scattered delivery data into one conversational surface, cutting the manual status-gathering that eats a manager's week. It is a working example of the agent and workflow-automation patterns ISEMI ships for clients: an Agent ADK router, MCP tool use across real business systems, and human-friendly answers where the team already works.