Five industries walked through — what AI-native looks like in the wild — step 4 of 9
Case 2: Insurance brokerage (commercial lines, small business)
(Illustrative scenario, not an actual case study.)
A 60-person commercial insurance brokerage. $12M revenue. Specializes in policies for small businesses (restaurants, contractors, retail stores). Renewal cycle every 12 months. The job is to gather the customer's data, compare quotes from 8-12 carriers, flag risk exposures, present options to the customer, place the policy, and service it for the year. Then do it again.
The work is 90% structured. Forms. Documents. Comparisons. Policy language. Premium calculations. Renewal deadlines. The 10% that isn't structured is the actual broker-customer conversation: "your business has changed, here's what that means for your coverage." That 10% is where the real value sits. The other 90% is where the margin goes to die.
The old workflow
Renewal is 90 days out. An account manager pulls up the customer's current policy. Emails the customer asking for updated information (payroll counts, square footage, prior claims, new business lines). Customer doesn't reply. Account manager follows up. Customer replies in fragments across three emails and a phone call. Account manager hand-fills the data into the carrier comparison tool. Tool spits out 8 quotes. Account manager compares them in a spreadsheet. Drafts a client memo. Sends to the broker for review. Broker tweaks. Broker meets with the customer. Customer asks about coverage gaps. Broker goes back to the carriers for clarification. Two more weeks pass. Policy gets placed three days before expiration. Account manager sets up the renewal calendar entry for next year. Account manager also services the policy mid-year when there's a claim, a coverage change, or a billing question.
A senior account manager handles maybe 80 accounts and works 55 hours a week. Most of those hours are NOT spent advising the customer. They are spent moving information between tools, remembering details, chasing carriers, and writing the same email in slightly different forms.
The AI-native version
The agent owns the data-gathering and the comparison. Specifically:
- Pre-renewal intake: 90 days out, the agent emails the customer with a structured questionnaire pre-filled with everything known about the business. Customer reviews and corrects. Agent flags changes (payroll up 30%, new business line) for the broker.
- Carrier outreach: agent submits the application to all relevant carriers automatically. Tracks responses. Flags questions back to the broker only when the carrier asks something a human needs to answer.
- Comparison memo: agent generates a structured comparison of the 8-12 quotes that come back. Highlights coverage gaps. Calls out cases where the cheapest option isn't the right one. Memo goes to the broker for review.
- Customer meeting prep: agent prepares a one-page brief for the broker covering changes year-over-year, gaps to discuss, upsell opportunities (cyber coverage just got cheap), and competitor activity in the customer's industry.
- Policy placement: once the customer picks a carrier, the agent files the paperwork, schedules the renewal calendar entry, and sets up the billing.
The broker does the customer meeting, the judgment calls on coverage tradeoffs, and the relationship work. The account manager becomes a supervisor of the agent — handling exceptions, training the agent on edge cases, and reviewing flagged decisions before they go out.
A 60-person brokerage becomes a 25-person brokerage running the same book. The senior brokers stop spending 30 hours a week on data movement and start spending 30 hours a week on customer relationships. Their close rate goes up. Their account count per broker doubles.
Where the wedge is
Two automations cascade through the whole business:
- Structured intake. If you can't get the customer's data into structured form without a human doing it, nothing else automates. The first investment is a renewal questionnaire that's pre-filled, validated, and a quarter as long as the carrier's official form.
- Quote-to-comparison generation. Once you have the data, the agent can submit it to all carriers and turn the 8-12 quote PDFs back into a structured comparison. This is the single most-time-consuming thing the account manager does. Eliminating it is the difference between a 10% margin and a 35% margin.
Everything else (policy service, billing, renewal scheduling) is downstream.
The metric that proves it worked
Three numbers:
- Time from intake start to client meeting: was 35 days (because nobody could find time). Target: 8 days.
- Account-manager-to-broker ratio: was 3:1 (account managers doing data work). Target: 1:2 (account managers supervising agents, more brokers because each is more productive).
- Customer retention rate: was 86% (customers churn when the renewal feels mechanical). Target: 94%+ (the customer meeting is now substantive, not transactional).
The same logic applies to legal services, accounting firms, compliance shops, and healthcare admin. Document-heavy, rule-based, deadline-driven, relationship-finished. Every one of those industries has the same 90/10 split between structured work and judgment. Every one is ripe.
What this case study teaches
- Document-heavy industries are the wedge. If a business runs on forms, contracts, comparisons, and structured comparisons, agents can swallow 80% of it. Insurance, mortgages, tax prep, HR compliance, accounting, legal intake — all the same shape.
- The customer experience changes shape. Customers in the old model talked to their broker once a year, the rest was emails about paperwork. In the AI-native model, the broker talks to the customer about substantive coverage decisions, and the paperwork is invisible. The customer feels more served, with less broker time.
- The account-manager role transforms, doesn't disappear. The old account manager was a data janitor. The new account manager is an agent supervisor. Same person, more leveraged work, higher compensation, lower headcount.
Five industries walked through — what AI-native looks like in the wild — step 4 of 9
Case 2: Insurance brokerage (commercial lines, small business)
(Illustrative scenario, not an actual case study.)
A 60-person commercial insurance brokerage. $12M revenue. Specializes in policies for small businesses (restaurants, contractors, retail stores). Renewal cycle every 12 months. The job is to gather the customer's data, compare quotes from 8-12 carriers, flag risk exposures, present options to the customer, place the policy, and service it for the year. Then do it again.
The work is 90% structured. Forms. Documents. Comparisons. Policy language. Premium calculations. Renewal deadlines. The 10% that isn't structured is the actual broker-customer conversation: "your business has changed, here's what that means for your coverage." That 10% is where the real value sits. The other 90% is where the margin goes to die.
The old workflow
Renewal is 90 days out. An account manager pulls up the customer's current policy. Emails the customer asking for updated information (payroll counts, square footage, prior claims, new business lines). Customer doesn't reply. Account manager follows up. Customer replies in fragments across three emails and a phone call. Account manager hand-fills the data into the carrier comparison tool. Tool spits out 8 quotes. Account manager compares them in a spreadsheet. Drafts a client memo. Sends to the broker for review. Broker tweaks. Broker meets with the customer. Customer asks about coverage gaps. Broker goes back to the carriers for clarification. Two more weeks pass. Policy gets placed three days before expiration. Account manager sets up the renewal calendar entry for next year. Account manager also services the policy mid-year when there's a claim, a coverage change, or a billing question.
A senior account manager handles maybe 80 accounts and works 55 hours a week. Most of those hours are NOT spent advising the customer. They are spent moving information between tools, remembering details, chasing carriers, and writing the same email in slightly different forms.
The AI-native version
The agent owns the data-gathering and the comparison. Specifically:
- Pre-renewal intake: 90 days out, the agent emails the customer with a structured questionnaire pre-filled with everything known about the business. Customer reviews and corrects. Agent flags changes (payroll up 30%, new business line) for the broker.
- Carrier outreach: agent submits the application to all relevant carriers automatically. Tracks responses. Flags questions back to the broker only when the carrier asks something a human needs to answer.
- Comparison memo: agent generates a structured comparison of the 8-12 quotes that come back. Highlights coverage gaps. Calls out cases where the cheapest option isn't the right one. Memo goes to the broker for review.
- Customer meeting prep: agent prepares a one-page brief for the broker covering changes year-over-year, gaps to discuss, upsell opportunities (cyber coverage just got cheap), and competitor activity in the customer's industry.
- Policy placement: once the customer picks a carrier, the agent files the paperwork, schedules the renewal calendar entry, and sets up the billing.
The broker does the customer meeting, the judgment calls on coverage tradeoffs, and the relationship work. The account manager becomes a supervisor of the agent — handling exceptions, training the agent on edge cases, and reviewing flagged decisions before they go out.
A 60-person brokerage becomes a 25-person brokerage running the same book. The senior brokers stop spending 30 hours a week on data movement and start spending 30 hours a week on customer relationships. Their close rate goes up. Their account count per broker doubles.
Where the wedge is
Two automations cascade through the whole business:
- Structured intake. If you can't get the customer's data into structured form without a human doing it, nothing else automates. The first investment is a renewal questionnaire that's pre-filled, validated, and a quarter as long as the carrier's official form.
- Quote-to-comparison generation. Once you have the data, the agent can submit it to all carriers and turn the 8-12 quote PDFs back into a structured comparison. This is the single most-time-consuming thing the account manager does. Eliminating it is the difference between a 10% margin and a 35% margin.
Everything else (policy service, billing, renewal scheduling) is downstream.
The metric that proves it worked
Three numbers:
- Time from intake start to client meeting: was 35 days (because nobody could find time). Target: 8 days.
- Account-manager-to-broker ratio: was 3:1 (account managers doing data work). Target: 1:2 (account managers supervising agents, more brokers because each is more productive).
- Customer retention rate: was 86% (customers churn when the renewal feels mechanical). Target: 94%+ (the customer meeting is now substantive, not transactional).
The same logic applies to legal services, accounting firms, compliance shops, and healthcare admin. Document-heavy, rule-based, deadline-driven, relationship-finished. Every one of those industries has the same 90/10 split between structured work and judgment. Every one is ripe.
What this case study teaches
- Document-heavy industries are the wedge. If a business runs on forms, contracts, comparisons, and structured comparisons, agents can swallow 80% of it. Insurance, mortgages, tax prep, HR compliance, accounting, legal intake — all the same shape.
- The customer experience changes shape. Customers in the old model talked to their broker once a year, the rest was emails about paperwork. In the AI-native model, the broker talks to the customer about substantive coverage decisions, and the paperwork is invisible. The customer feels more served, with less broker time.
- The account-manager role transforms, doesn't disappear. The old account manager was a data janitor. The new account manager is an agent supervisor. Same person, more leveraged work, higher compensation, lower headcount.