Senior Commercial Broker, Brokerage Intelligence
Harper · San Francisco
Senior Commercial Broker, Brokerage Intelligence
Role in one line: Turn expert commercial brokerage judgment into the systems, training, and product recommendations that define how Harper brokers insurance.
The hardest brokerage decisions rarely begin as clean rules. They live in the judgment of people who have seen a difficult risk, a thin submission, an unusual coverage structure, and a market exception before. Your job is to make that judgment teachable, testable, and reusable across Harper.
This role does not sell or originate insurance, own a book, carry a sales quota, close customers, or make final account-level placement decisions.
You will work deeply with product, engineering, operations, and insurance teams to build the brokerage infrastructure behind the work: the workflows people follow, the standards AI systems are evaluated against, the training that changes behavior, and the feedback loops that make the whole operation improve.
Harper is building an AI-native brokerage, and that only works if expert judgment becomes something people and systems can learn from. New York and San Francisco are the on-site homes for this senior brokerage and product work.
Why now
Harper serves many industries, coverage lines, and market paths. The company already generates enormous amounts of brokerage judgment through real accounts, market responses, exceptions, and outcomes. The opportunity is to turn that judgment into infrastructure instead of letting it disappear inside individual conversations.
The Senior Commercial Broker sits where commercial insurance practice meets product and strategy. You study how brokerage work is actually done, decide what excellent judgment looks like, encode it into workflows and decision standards, train AI systems and people against those standards, and use what fails in practice to improve Harper's products and operating model.
The mission is larger than documenting today's process. You will help define what the brokerage process should become when expert people and capable AI systems work together.
How you work with the team
You partner closely with product and engineering to turn brokerage judgment into useful tools and reliable AI evaluations. You bring the insurance standard and the evidence. Product leaders make roadmap and release decisions.
You also work with placement teams, licensed brokers, underwriters, and account owners who execute individual accounts. You define and evaluate how strong market selection, submission, negotiation, recommendation, and binding should work. They keep the final decision on each account.
You create the brokerage content, cases, and decision standards. Harper's learning and enablement teams help deliver them across the company.
What you own
Encode brokerage judgment. Turn complex commercial decisions into clear workflows, decision standards, playbooks, examples, and escalation rules without flattening away the judgment that matters.
Define brokerage product requirements. Work with product and engineering to identify the brokerage problems worth solving, define what the product must understand, and recommend capabilities grounded in expert evidence.
Train and evaluate AI systems. Build realistic cases, expected outcomes, grading standards, and failure examples. Review AI behavior against source evidence and expert judgment, then turn failures into permanent evaluations and product improvements.
Design and evaluate brokerage workflows. Encode how information, decisions, exceptions, and handoffs should move through account processes. Test whether the workflow produces sound judgment without taking over its individual accounts.
Own the brokerage learning standard. Create the content, cases, worked examples, decision standards, and effectiveness measures that make expert brokerage judgment teachable. Collaborate with the learning and enablement owners on delivery.
Run the expert feedback loop. Stay close to frontline work, difficult cases, market responses, and user behavior. Separate one-off exceptions from recurring patterns and route each to the right product, process, or training change.
Set evidence standards. Make sure recommendations and decisions remain traceable to the real application, document, communication, quote, or market response that supports them.
Raise the quality bar. Identify where a workflow produces inconsistent judgment, where an AI system sounds right but is wrong, and where a person cannot tell what to do next. Redesign the system until the standard is clear and usable.
What you will make possible
Priority brokerage workflows become clear enough to use, audit, and improve without flattening away expert judgment.
AI evaluations distinguish grounded commercial reasoning from an answer that merely sounds convincing.
Material failures become durable evaluation cases, product corrections, workflow changes, or explicit accepted limits.
Training built from live cases changes how people and systems handle the work.
Frontline evidence reaches product and engineering as a specific problem with a testable success condition.
Who thrives here
You have substantial commercial brokerage, wholesale brokerage, or placement experience across complex small and middle-market accounts.
You have built submission strategies, negotiated difficult terms, and seen enough real outcomes to know where written process ends and expert judgment begins.
You understand general liability, property, workers' compensation, commercial auto, umbrella, and specialty coverage well enough to recognize the material facts, tradeoffs, and exception paths.
You are excited to use your brokerage expertise to build products, systems, and standards rather than personally produce accounts.
You can turn tacit judgment into precise instructions, worked examples, and reviewable evaluations without pretending every decision is deterministic.
You can teach experienced people without talking past them, and you can translate insurance practice clearly for product and engineering partners.
You test your own assumptions against source evidence and change the standard when the evidence proves it wrong.
You are curious about AI, comfortable examining its failures in detail, and motivated by making expert work more consistent and scalable.
You are based in New York or San Francisco, or ready to relocate to one of those offices.
Who should not apply
You want your primary work to be personally producing and binding individual accounts.
You prefer keeping expert judgment in your own head instead of turning it into systems other people can use.
You treat documentation as a summary written after the real work rather than part of the operating infrastructure.
You want AI to sound convincing more than you want it to be correct and grounded.
You prefer a remote individual practice to an in-person, cross-functional build.
You are not willing to revisit a familiar brokerage habit when product evidence or operating outcomes challenge it.
Benefits
Harper offers health, dental, and vision insurance. Office-specific benefits are set by working location.
How to apply
Send your resume and a short example of expert judgment you made reusable. Describe a difficult commercial brokerage problem, what made the obvious rule incomplete, how you turned your judgment into a workflow, training tool, evaluation, or product improvement, and how you knew the new approach worked.