Want to know if this job is worth applying to?
San Francisco, California, United States
Onsite
Harper is an AI-native commercial insurance company in San Francisco. We're not bolting AI onto insurance — we're rebuilding the entire business as software, on a simple bet: turning expert human judgment into compute is one of the largest transitions left to make, and a trillion-dollar industry still run 90% by hand is the place to prove it. We've grown ~100x in the last year and we move at that speed — on-site, in person, long days, very high standards. Almost no one joins Harper for insurance; they join to build the company that replaces how it works.
A contractor needs a certificate before they can walk onto a job site. A payment failed. A customer thinks a policy change is already done when it is not. You pick up the call, establish what is actually true, and keep ownership visible until the request is complete. Small-business owners are used to a broker who never calls back. You are the reason ours do.
Harper is building service around a simple promise: the customer can always see what is true and who owns the next action. AI surfaces the account, preps the routine work, and drafts the follow-up. What is left is judgment on the phone — hearing the difference between confused and genuinely upset, catching the thing the customer did not say, knowing exactly who to pull in when the problem is bigger than the call — and the follow-through that closes the loop. That part is you.
You do not need an insurance background to start. We teach the systems, the playbooks, and the vocabulary. Quoting and coverage advice belong to licensed teammates, so you will never be asked to advise on a policy. Everything else, you carry as far as it goes. There are real paths from this seat into licensed service, sales, and operations.
Answer inbound calls with judgment. Pull up the account, understand the real request, and explain the next step in plain language. Tell the customer who owns the next action and when they will hear back, then keep that promise.
Take the outbound list to zero. Payments, documents, renewals, at-risk cancellations, and open service requests. Standard follow-ups get three attempts across three days; a payment issue gets up to seven touches by phone, text, and email. You stop when it is solved, not when it is hard.
Carry the issue forward. Complete the work you are authorized to do, and when a licensed or specialized teammate has to step in, hand off the request, the urgency, and the work already done so nobody restarts the conversation.
Recognize urgency. Hear the difference between a routine request and a problem that could interrupt coverage, operations, or a customer's ability to work, and escalate it before it does.
Keep the record clean. Clear notes, current status, named owner, next action, and the promised update point on every open request. The next teammate picks up mid-story and the customer never repeats themselves.
Protect the advice boundary. Coverage and quoting questions go to a licensed owner. You never guess.
Improve the service system. Flag repeat contacts, confusing communications, and broken handoffs so Harper fixes the cause instead of the symptom.
You have worked in a call center, dispatch, collections, front desk, customer service, or another high-volume phone environment.
You are clear, patient, and direct with people who may be confused or frustrated, and you can turn a rough call around.
You can manage a call, a record, and a next-action list at the same time — you talk and type simultaneously, and you can work a list of 80 calls without losing your place.
You are dependable and punctual, and you take that seriously.
You know when to solve, when to escalate, and when not to improvise. You would rather ask than guess.
You treat AI as leverage — it finds the account and preps the routine work so you can spend the call on the customer — and you have opinions about where it helps and where it gets in the way.
You learn systems and insurance vocabulary quickly. Spanish fluency is valuable.
This job is the phone. On-site in Atlanta or San Francisco, Monday through Friday, in the building with the team. Two shifts cover the day — an early one for East Coast customers and a later one through the West Coast close — and you are assigned to one based on business need. Overtime is available and the volume is real: back-to-back inbound calls and long outbound lists, all shift, with an empty list at the end of it as the whole game.
This is not a low-volume relationship-management role, and it is not the seat for someone who thinks a transfer counts as resolving the problem, who avoids outbound follow-up or difficult conversations, who needs a remote or hybrid default, or who is comfortable giving an answer they are not sure is right. It is for someone who likes talking to people all day and ends it energized, who is bothered by an open item until it is closed, and who wants a first seat at an AI-native company that is actually shipping. If the intensity reads as a cost you'd rather avoid, this isn't the seat. If it reads as the point, keep going.
Salary: 23–26/hour, full-time, non-exempt, overtime eligible.
Equity: Yes.
Location: On-site, Atlanta or San Francisco. Based there or willing to relocate.
Schedule: Monday–Friday, in-office. Early shift (East Coast coverage) or late shift (West Coast coverage), assigned based on business need. Overtime available.
Health, dental, and vision insurance
Meals provided — breakfast, lunch, and dinner
Snacks, drinks, and coffee stocked daily
Commuter benefits
Free gym membership
Team call — 30 minutes on Zoom. You meet the team, we meet you.
Office session — about two hours in person. You meet more of the team and work a real problem with us.
Super day — how you actually operate, in real time. We decide after this.
The process moves as fast as you do.
Send your resume and tell us about the busiest customer-service environment where you worked — the kind of requests you handled, how you kept follow-through clean, and one time you recognized that a routine call was more urgent than it first sounded.