Scaling onboarding is deceptively hard. As Abnormal rolled out R&D Onboarding 2.0, the goal was clear: new hires should have access to AI tools on day one and be shipping code within their first week. Achieving that required a mix of facilitated sessions and self-serve training across tools like Claude Code, Cursor, and Nora.

For Chanel Green, a Technical Program Manager focused on AI enablement, that vision created a new operational challenge.

Every Friday, Chanel reviewed a spreadsheet of upcoming new hires and manually added each person to multiple onboarding sessions on Google Calendar. At small scale, the process worked. As hiring increased, it quickly became time-consuming and error-prone, especially during weeks when seven to ten new engineers started at once.

Chanel explored existing tools, but none were a good fit. Glean lacked direct Google Calendar integration, and using Nora would have required additional APIs and coordination with other teams. What she needed was something lightweight, fast, and fully self-serve.

Chanel partnered with Claude Code to build a solution tailored exactly to Abnormal’s onboarding workflow. She used Google Apps Script to create an automation that runs directly inside the onboarding spreadsheet, using her existing permissions.

The scheduler adds a custom menu to the spreadsheet, allowing her to run a dry run first. This preview shows which new hires will be scheduled and which sessions they’ll be added to, making it easy to catch issues before sending invites.

Once confirmed, the scheduler automatically adds eligible new hires to the correct calendar events. If a hire doesn’t yet have an email address, they’re skipped safely until the information is available. What used to require careful manual coordination now runs with just a few clicks.

Claude Code played a critical role in helping Chanel iterate. She used it conversationally to troubleshoot edge cases, refine logic, and validate the automation step by step, learning along the way without needing deep scripting expertise.

The onboarding scheduler doesn’t just save time. It establishes a repeatable pattern for identifying low-leverage work and automating it with AI.

By building the solution herself, Chanel demonstrated how TPMs can own their own automations rather than waiting on centralized tooling. The process also builds practical AI fluency: learning how to reason through workflows, test assumptions, and iterate quickly using Claude Code.

What started as one automation has already led to several more.

The onboarding scheduler now scales cleanly across North America and London, with plans to extend to APAC. It removes a fragile dependency on a single person and ensures onboarding continues smoothly even when Chanel is out.

More broadly, it unlocks leverage for the TPM team. Starting next week, TPMs will be enabled with Claude Code using a guide Chanel created, allowing them to adapt this pattern to their own workflows while building hands-on AI skills.

The next step is full automation. Chanel plans to have the scheduler run automatically every Friday without manual triggers. Future enhancements include manager notifications and personalized onboarding guides tailored to each new hire’s division.

Longer term, this work sets the foundation for a culture where TPMs regularly identify manual processes, automate them with AI, and compound efficiency over time.

The onboarding scheduler is just the beginning, but it’s already proving how powerful self-serve AI enablement can be.