Our agents talk to people about their debts, their bills and their loans. A wrong sentence can hurt someone or break a law. If you want your work to be judged by what happens on real calls, not by a slide, you will like it here.
Most AI companies learn about contact centers from their customers. We have run them since 1955. That changes what you build and how quickly you find out whether it works.
You work on collections, lending, billing and service calls where identity checks, disclosures and consent are not optional. The problems are hard in a way that matters.
Agent Studio, simulation and ramp gates let a change reach a small share of live traffic safely, so you see the effect of your work on real outcomes.
Write gates, audit chains and evaluation suites are first class engineering, not paperwork. Doing it properly is the job.
An agent that works on a live call needs engineering, language, quality and operations. These are the teams that make it happen.
We would rather see how you think about a real problem than how you do on a puzzle.
Send your CV and a few lines on the team that interests you. A person reads every message.
We explain the work in plain terms and you ask us anything, including the hard questions.
A realistic task close to the role, such as reviewing a call flow or scoring a transcript. No trick questions.
We tell you where you stand and why, whichever way it goes.
We do not publish every role on this site. If you want to work on governed AI for customer conversations, email hello@billgosling.ai with the subject line “Careers” and tell us which team you would join and where you are based.
The fastest way to understand the job is to watch an agent take a call, hand over to a person and leave a signed record behind.