Buy Insurance Agencies, Add AI, Collect Renewals: The $70M Playbook

American Growth Insurance (AGI) launched on 16 July 2026 with nearly $70 million in committed equity to buy independent insurance agencies and rebuild them around AI. The company says that before buying anything, it worked with 10 agencies for about a year and lifted their average profitability by more than 50%. That is a company-reported figure, and it is the whole thesis: buy a book of renewals, cut the admin cost of servicing it, keep the clients.

You do not need $70 million to test a smaller version. This post covers how the model works, the deal math at three sizes, who else is doing it, and what a first-time buyer can realistically copy.

What is the insurance agency roll-up opportunity?

Independent agencies are numerous, small and often have no successor. DealSeam, citing OPTIS Partners, says over 30,000 independent agencies have under $1.25 million in revenue, most with no ability to perpetuate. The same page puts the count of US independent agencies at about 37,000 in 2026, down from about 39,000 in 2024.

AGI's launch release cites MarshBerry data that the top 50 US brokerage firms hold 96% of broker revenue, with the next 50 holding 4%. Atomic, AGI's venture studio backer, says it reviewed more than 1,000 service industries before choosing this one.

The "why now" is AI. Agents are already using it individually. Liberty Mutual's 2026 study found 65% of agents used AI for work in the past year, up from 37% in 2025, and agents who use it say it saves about four hours a week. But only 14% of agencies had rolled out an AI tool formally. That gap is what an acquirer claims it can close.

How much can an insurance agency make?

It depends on the price you pay and the debt you carry. Small owner-operated agencies typically sell for about 2.9x to 4.3x SDE (seller's discretionary earnings, roughly owner pay plus profit), and tracked EBITDA multiples run from about 4.3x to 11x, according to DealSeam as of 6 October 2026. Treat these as a starting point for negotiation, not a quote.

Below is the math at three sizes. Every figure is an example estimate, not a forecast or a promise.

Assumptions: earnings hold flat after the sale, no client loss, no AI savings, no income tax. In the small and medium cases the owner's salary is paid out of SDE. The large case uses EBITDA, so management salaries are already deducted. Interest rates are assumptions, not quotes. The medium loan may exceed what some lenders will fund on one agency, so confirm with a lender.

The honest read: a small agency can replace a salary, not create wealth, once debt is paid. If a buyer loses 10% of the earnings in year one, the small case drops from about $84,400 to about $69,400 before tax. AI helps only if it cuts costs or lifts revenue by more than the tools cost. AGI's 50% claim is not a number to plan around.

Which companies are already doing this?

Three funded companies are close to this model, and two others show adjacent paths. All funding figures below are as announced by the companies or reported by trade press.

AGI. Its first acquisition was Heller-Kowitz Insurance Advisors in Baltimore, announced 6 August 2026. Terms were not disclosed. Heller-Kowitz was founded in 2014 and writes personal, commercial, employee benefits and life business. AGI says it targets more than $10 million in revenue by the end of 2026.

Gyde. After launch, it acquired Benavest, a national health agency focused on individual ACA, Medicare and ICHRA plans, announced 20 April 2026.

Harper. The company says it served more than 5,000 businesses in 13 months, placing with more than 165 carriers. The company says it operates as a brokerage rather than selling software to brokers. That makes it a competitor to agencies, not a roll-up.

What did they do with millions, and what can you do with $5,000 to $50,000?

The funded companies are buying scale and building proprietary software. Their capital pays for engineers, acquisition cheques and the years before profit. You cannot copy that with $50,000, but you can copy the order of operations.

AGI tested its operating model on 10 agencies before it bought one. With a small budget, run the same experiment on a single agency you work in, advise or plan to buy: pick one workflow, measure hours before and after, then decide. For a purchase, one bank advertises as little as 10% down on a first agency under its SBA program. Terms vary by lender, and lenders will look at your experience.

The same model works outside insurance. See how it plays out in buying a retiring CPA's firm and in owning a plumbing company without being a plumber.

Do you need a licence to buy and run an insurance agency?

Someone in the business must be licensed, and rules are set state by state. NIPR says an insurance producer licence allows a person to sell, solicit or negotiate insurance, and that producers must hold a valid licence in every state where they do business.

• Resident licence: for the state where you live. NIPR says it requires state-specific pre-licensing education and passing the exam.
• Non-resident licence: lets you sell in other states, applied for through NIPR with proof of your resident licence.
• Other lines: surplus lines broker, adjuster and third-party administrator licences are separate, and states differ.

Licensed work needs trained, licensed people. Before you sign, have a lawyer and your state insurance department confirm what the purchased entity, its staff and you personally must hold.

How much does it cost to start?

Cash at close is mostly the down payment plus a reserve for debt service. The table uses the small and medium cases above, so the same assumptions apply.

These are planning assumptions, not quotes. The 10% down figure comes from one lender's marketing, so get written terms. The typical column keeps the same reserve for simplicity, though a 20% down payment would shrink the loan and the reserve. A $50,000 budget does not cover even the low small-agency case unless the seller finances part of the price.

How do you buy an insurance agency, step by step?

1. Get licensed in your resident state through NIPR's state pages, and keep continuing education current.
2. Work or shadow inside an agency to learn the management system, carrier portals and renewal workflow.
3. Pick a niche. AGI spans personal and commercial lines; Gyde chose health. Specialising makes targeting sellers easier.
4. Source sellers through brokers, state agent associations and your own network. DealSeam, citing OPTIS, says over 30,000 small agencies have limited succession options.
5. Request retention, carrier mix and book data before any offer, and have an accountant check them against tax returns.
6. Value the agency with your lawyer and an accountant, using multiples only as a guide.
7. Line up financing with a lender experienced in agency loans. Ask about seller notes.
8. Negotiate protections: a transition period, non-solicit terms and an earn-out tied to retained revenue.
9. Confirm carrier consents and licensing before closing.
10. Close, then avoid big changes for 90 days. Keep staff and service levels steady.
11. Add AI one workflow at a time, and track hours saved.
12. Re-measure retention at 6 and 12 months before buying a second agency.

Which free tools and resources can you use?

Free chatbots carry data risk. In Liberty Mutual's 2026 study, as reported by Insurance Business, only 22% of respondents trusted AI with client and business data, and 18% said their agency had a well-defined AI policy. Write a policy before you paste a client file into a free tool.

How can you use AI in an insurance agency?

Start with the repetitive work. Pulse 2.0's summary of AGI's plans lists account servicing, data entry, policy comparisons, sales support, client communications and internal knowledge management as the areas where AI could be used.

Harper shows the commercial version. The company says its models read insurance applications, route submissions, follow up with underwriters and manage client communications, and that it gets businesses covered in 24 to 48 hours instead of weeks. Those are company claims.

Renewals are the natural first target. Rival Insurance Technology's Eric Purdy describes a "system of work" that notices a renewal is due, gathers what changed and helps the broker decide what to do next, per Insurance Business. Pricing for agency AI tools is mostly not published, so get quotes. If you would rather sell the software than run the agency, selling AI to accountants shows a vendor-side version of the same idea.

How do you scale it, and what are the risks?

Scaling means a second agency, and the constraint is integration, not capital. Purdy calls swapping systems a "change management problem". The Big "I" survey cited by Insurance Business found only 8% of agencies have AI embedded in daily workflows, and 55% have no written AI policy.

Free tools stop working when client data, carrier portals and multi-user permissions matter. At that point you need agency management software and an AI policy.

Business failure is the base rate to respect. BLS cohort data, as summarised by StartBusinessByState, shows 51.4% of employer establishments opened in March 2020 were still reporting five years later. Workspace369 shows 50.2% of the 2015 cohort survived five years. These are all-industry figures, and a closure is not always a failure.

Other risks: client attrition after the sale, carrier terms, key-producer departure, errors and omissions claims and state licensing. This is research, not legal, tax or financial advice.

What's still unclear?

• The headline figure. AGI's $70 million is "committed equity," not cash raised. Some trade headlines say "raised."
• Whether AI works at scale. The more than 50% profit lift comes from AGI, across 10 agencies, with no independent audit.
• Purchase prices. AGI and Gyde did not disclose deal terms, so the multiples AI-native buyers actually pay are unknown.
• Multiples. The 2.9x to 4.3x SDE range comes from one aggregator, DealSeam, summarising published ranges.
• Market direction. OPTIS counted 292 deals in H1 2026, the slowest start since 2016 according to Axia Growth.
• Weakest point. An agency's value is its client relationships. If AI makes service feel less personal, the book you paid for can walk.

For a lower-capital angle, see AI bookkeeping for small businesses.