Case study1 min read

Building a Mentorship Program From the Ground Up

How we matched 300 mentors and mentees using an internal AI pairing tool, then built the rollout around it

The situation

There was no mentorship program to inherit, no matching process, no shared understanding of what a good mentor-mentee relationship even looked like.

Pairing 300 people well isn't something you can do by hand or by guesswork, and even a good match means nothing without the surrounding structure to help both sides actually use it.

What I did

Pairing at scale with the right tool for the job

We used an internal AI tool, built by our engineering team, to pair mentors and mentees based on what would actually make a good match, not just availability. That let us pair 300 people thoughtfully instead of manually, at a scale that manual matching never could have handled well.

Building the structure around the match

A good pairing tool doesn't run a program by itself. Alongside the matching, we built the communications and rollout plan to launch it, and self-service pages so anyone in the program could look up what a strong mentor-mentee relationship actually looks like, expectations, cadence, how to get the most out of it, without needing to ask.

The result

A mentorship program that didn't just launch, it had structure behind it: a scalable way to match people well, and the self-service resources to help both sides actually make it work.

Some of the program details are proprietary, but the short video below captures the gist of it.

Walkthrough of the mentorship program experience.