Editor, Agent Experience

About the role

STLabs is hiring someone to own how our agent comes across to the people it helps. The agent resolves IT and HR requests in Slack and email for employees who are often stuck or irritated; whether it feels competent, calm and helpful in those moments is the job. You own the agent's voice and the standard for a good interaction. Engineers build the models, prompts and evaluation tooling; you decide what they aim at and show them, with examples, when they miss.

We are not hiring from the usual AI/ Tech pool. We welcome candidates from hospitality, patient experience, editorial, teaching, support quality, and help-desk leadership. A degree in literature, philosophy or the humanities from a top school is required; beyond that, we care about evidence you have held others to a standard, not further credentials. ITSM or AI experience is useful and not required.

As an equal opportunity employer, we don’t tolerate discrimination or harassment of any kind. Whether that’s based on race, ethnicity, age, gender identity, citizenship, religion, sexual orientation, disability, pregnancy, veteran status or any other protected characteristic as outlined by federal, state or local laws. The reasonably estimated yearly salary for this role at is: $100,000—170,000 USD

Responsibilities

Write and maintain the agent's voice and behavioral guidelines: how it explains, asks, admits uncertainty, declines, escalates and closes a loop.; Read real transcripts every week and turn what you find into precise, example-backed asks for engineering.; Own the rubric for what a good interaction is, engineers run the evals, you define what passes.; Rewrite the agent's worst answers into the ones it should have given, and build the exemplar library the team learns from.; Decide handoff behavior: when the agent acts, asks, or stops and routes to a human.

Requirements

An ear. You notice when something is technically correct but lands badly, and you can say why.; Unusually good plain-English writing for readers who are busy and not in a good mood. Most of the output is written guidance and rewrites.; You have held other people to a standard across a high volume of real interactions: ran a desk, edited a team, trained staff, or graded conversations against a rubric.; You can read thousands of transcripts, keep the patterns in your head, and turn a pile of cases into a rule that is specific and hard to game.; Comfortable with a product that changes weekly. Your guidelines ship Friday and we find out what breaks.