Sit with a partner team (editorial, sales, production, ops - it rotates) and figure out where their week actually leaks: which tasks repeat, which handoffs break, which decisions get re-made over and over.

Turn what you find into something the tech team can build. That means user flows, sample inputs and outputs, edge cases, and the success criteria for “did this agent actually help.”

Write the agent briefs and skills documentation — short, concrete, testable. Not 20-page PRDs.

Run the rollout: onboard the first users, sit next to them while they use it, collect what works and what breaks, feed it back into the next iteration.

Help maintain the internal knowledge base (we call it the Brain) so the agents have something real to work from.

Track usage and outcomes. If an agent isn’t being used, we want to know within a week, not a quarter.

You can write clearly. If you can’t explain a workflow in a paragraph that someone in sales can act on, the rest doesn’t matter.

You’re curious about how things actually work - including the unglamorous parts (who approves what, where the spreadsheet lives, why nobody updates it).

You’re comfortable around AI tools. You’ve poked at ChatGPT, Claude, or similar enough to have opinions about what they’re good and bad at.

You can hold a conversation in both English and Vietnamese. Most of our work is bilingual.

You’re studying business, communications, design, CS, or something adjacent - or you’re self-taught and can show us what you’ve made.

Bonus: you’ve shipped something before. A side project, a club initiative, a small business. We want to see that you finish things.