An employee at GitHub automated marketing event management using GitHub Copilot
An employee at GitHub described how, through a conversation with GitHub Copilot (without writing code), they turned a runbook for preparing marketing events into an automated process using GitHub Issues and GitHub Actions that manages the entire event lifecycle.
The person responsible for marketing at GitHub in Japan and Korea described how they turned the recurring preparation of company events (webinars, meetups, executive meetings) into an automated process. Instead of writing code, they wrote a runbook as a markdown file (AGENTS.md) containing rules for naming campaigns, mapping fiscal quarters to dates, time zones for individual regions, and invitation templates. They shared this runbook in a conversation with GitHub Copilot, which gradually built a complete automated process based on it.
The resulting system works as follows: based on the conversation, GitHub Copilot proposes a campaign name, an invitation draft, and dates, which the author approves before launch. Creating a GitHub Issue with the label "event-setup" triggers a GitHub Actions workflow that completes steps in a few minutes that previously took almost an entire day. According to the author, a daily task triggered by a scheduler (cron) also runs every morning to check newly registered participants and, for invitation-only events, screen people on the waiting list against specified criteria; the system is also supposed to handle cleanup after the event on its own.
The author argues that off-the-shelf marketing automation platforms would not accommodate the differences between markets in the APAC region (different languages, different segmentation, different CRM fields by subregion), while this approach allows a workflow change to be introduced as a pull request that goes through standard review. According to the description, the conversation with Copilot originally took place in the terminal through GitHub Copilot CLI; more recently, it has also become available in the GitHub Copilot desktop application, lowering the entry barrier from "using a terminal" to "being able to type text". The author emphasizes that the principle (write a runbook, give it to a conversational AI agent, let it build an automated process) applies generally to any tool with scriptable access through an API or CLI, not just event management.
A description of the final part of the process (specific criteria for screening people on the waiting list and further steps) is not available in the source. Details can be found in the source article.
Why it matters
The case illustrates a concrete pattern for using a conversational AI agent for recurring nontechnical processes: a person defines rules in a text runbook without needing to write code, and the agent builds working automation from them using existing tools (GitHub Issues, Actions) as well as external systems with an API or CLI. For businesses, this suggests an alternative to purchasing specialized automation platforms where processes are fragmented by region or segment and customizing an off-the-shelf tool would be costly.
Two audiences, two different impacts
What this means
For individuals
According to the described approach, anyone handling recurring tasks in a tool that offers an API or at least a CLI can write their own runbook and let a conversational AI agent (GitHub Copilot) build an automated process from it, without needing to know how to write code.
For a business
The marketing team at GitHub replaced manual event preparation that took several days (registration, tracking sign-ups, post-event cleanup) with an automated process managed through GitHub Issues and GitHub Actions, which it launched through a conversation with GitHub Copilot instead of writing code; according to the author, this approach is more flexible than deploying an off-the-shelf marketing automation…
Sales and marketingCheck the original
Event sources
only one source so far · 1 publisher, 0 independent. We count feeds from the same owner only once.