Harper's AI Insurance Brokerage: What a Solo Agent Can Copy

Harper, which describes itself as an AI-native commercial insurance agency, raised $46.8 million in combined seed and Series A funding. The round was announced in late February 2026 and led by Emergence Capital. The company says it serves more than 5,000 customers. For a solo agent, the useful question is which parts of that model need venture capital and which parts can be copied with a producer license and AI software that already exists.

How much did Harper raise, and who led it?

Harper raised $46.8 million in combined seed and Series A funding, according to The Daily Star, which reported the round on 1 March 2026. A daily.dev post citing TechCrunch dated the announcement 25 February 2026 and also named Emergence Capital as lead. Some outlets report $47 million. This post uses $46.8 million, the figure in most reports.

Harper was founded by CEO Dakotah Rice and CTO Tushar Nair, a former Poolit CTO, and launched in 2024, according to The Daily Star. The company matches small and mid-sized businesses with more than 160 insurance carriers. That carrier count is the company's own figure, and it is not independently checked here.

What does Harper's AI actually do?

Harper says its AI routes submissions, handles underwriter follow-ups, collects documents and manages the pipeline. Rice described these manual steps to TechCrunch, as reported by The Daily Star. The company's stated aim is to automate them, which is a claim about design, not a measured result.

There is no independent measure of Harper's speed, accuracy or client outcomes. Every operational claim comes from the company or from outlets repeating it. That matters because the model's whole argument is that AI makes an agency faster and cheaper to run. A reader should treat that as the thesis to test, not a finding.

Who sells AI to agencies instead of running one?

Harper runs its own agency. Applied Systems and Vertafore sell software to agencies and brokerages, and both now sell AI as an extension of that software. The difference between the three is who carries the risk and the customer relationship.

Harper is betting on a full-stack agency. Applied and Vertafore are betting that agencies will buy AI inside the software they already run. For a solo agent, the second path is the practical one. Ask each vendor which AI features are live in the plan you would buy, and what they cost. Vendor pricing varies by plan and is not covered here.

What can a solo agent copy from Harper?

Harper's scale shows up in third-party figures. Startup Intros, a company profile site, reports more than $6 million in annualized premiums across 35 states and more than 5,000 businesses (Startup Intros). Both numbers are "over" figures, so they are floors, not exact totals.

Premium is not income. An agency keeps a commission on premium, and commission rates are not covered here, so no agent earnings figure is given.

Harper's funding is large relative to the premium it manages, which is why its model needs capital before it scales. The table sorts the parts of the model by whether they need that capital.

The copyable part is the workflow, not the scale. A solo agent who uses one agency system with AI features, works a small set of carriers, and writes in one state is running the same kind of operation at a smaller size. Whether that pays depends on commission rates and client volume.

Does a solo agent need a license, and what do the AI rules require?

A solo commercial agent needs a producer license from their resident state. Fees, exam rules and timelines differ by state, so get the license cost from your state insurance department before budgeting.

The NAIC adopted its Model Bulletin on the Use of Artificial Intelligence Systems by Insurers in December 2023. It states that existing insurance laws apply to AI systems and sets expectations for governance, documentation, testing and third-party oversight (NAIC issue brief). The bulletin applies to all authorized insurers (Regulations.ai). Whether a solo agency falls under it depends on how each state applies it. Check with your state insurance department.

The practical steps are the same either way: keep a record of which AI tools touch client data, test outputs before they reach a client, and keep a licensed person responsible for every recommendation. These are sound practices. They are not a legal checklist.

What's still unclear?

• Funding figure. Most reports give $46.8 million, and others say $47 million. The Silicon Valley Post lists Pioneer Fund as a backer. Other reports leave it out.
• Carrier count. The Daily Star says more than 160 carriers. FinTech Global cites more than 165 underwriters. The gap is unexplained.
• Company figures. The 5,000 customers and $6 million in premium are company or database figures. Startup Intros is a single source, and both numbers are "over" values, so they are floors, not exact totals.
• Performance. No independent measure exists of how fast Harper places coverage or how accurate its AI is.
• Weakest point of the argument. The case that a solo agent can copy the model rests on vendor features whose prices and fit for a solo agency are unknown, and on an NAIC bulletin whose reach to agencies is unsettled.

This is research, not legal, tax or financial advice.