The Problem

36 million businesses in America need insurance—it's not optional. 77% are underinsured. 40% have no coverage at all. The distribution system failed them: too slow, too opaque, too confusing.

Over 90% of commercial insurance is still human-led. We're building the inverse: 90%+ AI-led, pushing toward the higher 90s. Not by patching legacy workflows—by building AI that makes humans more effective, improves the customer experience, and eliminates friction at every step.

We're adding ~1,000 customers per month. We've grown 100x since last year. We're looking to do even more this year—and that's why we're hiring.

The Thesis

Every industry with human-bounded distribution consolidates rapidly once someone makes it computational. Search before Google. Ride-hailing before Uber. When distribution becomes computational, Jevons Paradox kicks in: efficiency leads to expansion. When getting the right coverage becomes fast and frictionless, the 77% of underinsured businesses will finally get properly protected.

The companies that win this transition won't just have great AI. They'll have figured out how to organize themselves around it—how knowledge gets encoded into systems queryable by agents and operators. The organization is the moat. We're at the forefront of figuring out what an AI-native company looks like, and this role sits inside that question.

Harper isn't an AI tool sold to brokers. We are the broker. We do the work end-to-end and sell the outcome: small businesses get the right coverage, fast, at the right price, with the right service. Same-day quotes, instant certificates, real answers, 24/7. Owning both sides is the moat.

The Role

Harper operates like a factory with a series of modules spanning the full lifecycle from intake through renewals. Across them we run a stack of internal AI systems covering operator guidance, the operational backbone that matches risks to underwriters, autonomous communications, and voice AI for customer interactions.

You'll own a module inside this factory end-to-end: the customer experience, the operator workflows, the AI agents underneath. You run it with an FDE and the operators who live inside it: sales, service, underwriting, ops. You decide what to build and, more importantly, what not to build.

Our founding PM team has gone wide across every module. We need to go deep. Industry by industry, segment by segment, edge case by edge case. How a daycare buys insurance vs how a trucking company buys it. What "urgency" means for a tow yard with expiring dealer plates vs a GL renewal. Why loss runs matter more for some businesses than others. We need to encode that nuance into the systems so they do the work as well as a human in most places, and better than a human in many.

We're hiring early-career PMs on purpose. As we expand, you'll move modules. You'll get the kind of surface area and reps that take a decade somewhere else.

What You'll Do

• Own the KPIs — Conversion, handle time, accuracy, autonomous resolution rate, retention—whatever the leverage point is for your surface. You set the targets, instrument them, move them. If the metric isn't moving, that's your problem.
• Encode the nuance — Every module has industry-by-industry, segment-by-segment behavior to encode. You translate what makes your module's customers different into rules, prompts, agents, and data structures.
• Own the eval regime — Probabilistic systems are only valuable when humans trust them. Regressions on every change, evals that map to real outcomes (not vibes), backtests against historical applications, call-by-call review where it matters. You'll be paranoid about silent regressions in a way most PMs aren't.
• Build the data flywheel — Work hand-in-glove with data labeling and validation to build the golden datasets your module's models need. You define what "right" looks like so we can train, evaluate, and improve against a real bar.
• Own the cross-modal experience — Your module spans web, voice, and human. You decide where each modality wins, where they hand off, and how the on-ramps feel.
• Live with operators — Sit with sales, service, underwriting. Watch the work. Find what's broken before they tell you.
• Talk to customers every day — Literally. Not "5 calls last quarter."
• Prototype with AI — Claude Code, Cursor, Lovable. Walk into the meeting with a working prototype, not a deck.
• Hyper-prioritize — Out of 50 things people are asking your module to do, find the 3 that move the KPI and ignore the rest with conviction.

You Might Be a Fit If...

• You get what an AI services company is. We're not selling software. We're doing the work and selling the outcome. You're shipping behavior into a probabilistic system that real operators and real customers have to trust.
• You're obsessed with evals. You'd rather ship a worse model with a great eval harness than the other way around.
• You think in KPIs—not "we shipped the feature" but "we cut handle time 40%." If you can't tie what you build to a number, you don't ship it.
• You can sit with operators. You'll go sit with the underwriting team for a week and come back with the 5 things that matter.
• You can build. Cursor, Claude Code, Lovable. You'd rather show the thing than describe it.
• You can talk shop on AI—agents, LLMs, context engineering, data pipelines, evals. You don't write the code, but you can argue tradeoffs with the engineer who does.
• You go deep before going wide. You enjoy learning one industry well enough to encode its quirks. You don't get bored two layers in.
• You want to own a thing, not coordinate a thing. If "PM" sounds like meetings to you, this isn't it.

Requirements

• 1–3 years into product, OR an early-career operator, engineer, or AI researcher who's been doing the work without the title
• Demonstrated ownership of a product or system end-to-end—KPIs, roadmap, execution
• Proficiency using AI tools to prototype (Claude Code, Cursor, Lovable, or similar)
• Strong analytical instincts—you can argue tradeoffs with engineers on AI systems
• Track record of going deep on a domain and encoding what you learned into a system
• Based in San Francisco or willing to relocate

Nice to Have

• Background in AI/ML products, voice AI, agent frameworks, or workflow automation
• Experience with eval design, prompt engineering, or context engineering
• Insurance, fintech, or regulated industry experience
• Prior startup experience

Compensation

Salary: $125,000–$170,000 + performance bonuses & equity

Location: San Francisco, in-office

Schedule: Monday–Friday, 5 AM – 8 PM. The hours are long. The learning curve is steep. The people who thrive here wouldn't have it any other way.

Benefits

• Health, dental, and vision insurance
• Commuter benefits
• Team meals and snacks

The Process

20-min lead screen — Alignment on mission and pace

Technical conversation — Walk us through analysis you’ve done

On-site — Meet the team, see the data, show us what you’ve got

To Apply

Most PM jobs are about navigating someone else's roadmap inside someone else's category. This one is different. Commercial insurance has never been rebuilt. The gap between what frontier models make possible and what's actually shipped in this industry is the largest it will ever be.

If you want to own a module that touches real customers from week one, work alongside founders and engineers who ship every day, and learn how to build, measure, and trust AI systems at production scale—send your resume and tell us about something you built that moved a number.