Brown & Brown — Choosing the First AI Workflows to Scale
Download case PDF (5 pages)Request teacher’s pack
Decision date and information cutoff: 9 September 2026, 23:59 Eastern Time
Decision scope: First scaled workflow within Brown & Brown’s announced AI-first transformation
Case convention: This is an author-created management decision exercise. The company’s announced AI direction and operating model are historical disclosures. The workflow alternatives and assignment are instructional constructs; no private deliberations are represented.
The decision
On 23 July 2026, Brown & Brown announced that it was moving toward an AI-first enterprise. Its plan included deploying Anthropic’s Claude across approximately 23,000 teammates, working with Anthropic, McKinsey and Accenture, and establishing a value management office (VMO) to monitor adoption, impact, return on investment and controls. The company said it would develop solutions close to local businesses and customers while building an enterprise platform for wider adoption. [S1]
The first disclosed results came from selected software-engineering teams using Claude Code. Participating teams reported productivity gains of approximately 2x–8x in some development work, 80–90% faster analysis or troubleshooting in certain use cases, and high confidence in the tools. Brown & Brown did not disclose the underlying task mix, sample, comparison group or cost conversion. The release identified these early results and expected benefits as forward-looking statements. They did not establish results across the brokerage business. [S1]
The September company overview described Brown & Brown’s decentralized sales and service model: certain functions were centralized, while sales and service decisions remained local. It also identified technology, data and AI as enterprise investment areas. [S2]
Meanwhile, Q2 revenue grew 30.4%, largely in a period of substantial acquisition contribution, while organic revenue declined 0.7%. Performance varied by segment. Students must decide how the company should sequence the next stage of its announced program.
Assignment: Recommend one workflow for Brown & Brown’s first scaled AI wave: (1) Retail service and renewal preparation, (2) Specialty Distribution submission and placement support, or (3) a common internal support process. Specify the target users, what the AI may and may not do, how local teams and the VMO should share responsibility, and what evidence would trigger expansion, redesign or a stop. These workflow options are author-created; the company has not publicly selected them. Use only information available by the cutoff.
A brokerage built on local relationships and specialist knowledge
Brown & Brown distributes insurance and related services through two reported segments. Retail connects individuals and businesses with insurance products and services. Specialty Distribution includes specialist businesses, insurance programs and wholesale brokerage that works with independent agents and brokers. Its proposition depends on matching client needs with available insurer coverage and on the expertise of its local and specialist teams. [S3]
The company’s September overview reported 2025 revenue mix of 59% Retail and 41% Specialty Distribution, more than 700 locations, approximately 23,000 teammates and operations in 16 countries. These company-reported measures convey the scale at which the new technology direction could operate. [S2]
Brown & Brown describes its operating model as decentralized: local teams make sales and service decisions, while selected central functions support operating efficiency. Its company overview presents an entrepreneurial culture as a core principle. In July, management said local teams would incubate solutions close to business and customer needs, with enterprise capabilities to support broader adoption. [S1–S2]
This model creates a strategic design problem. A tool built locally may fit a workflow and its data, but remain difficult to govern or reproduce. A common enterprise platform may improve reuse and oversight, but require local teams to adapt distinctive customer and carrier workflows to shared standards. Brown & Brown’s public plan combines both levels; the open question in this case is which workflow should receive the first scaled wave and how evidence should guide it.
Revenue growth with different underlying trends
Brown & Brown reported Q2 2026 total revenue of $1.676 billion, up $391 million or 30.4% from the prior-year quarter. It reported Q2 organic revenue of $1.176 billion, down $8 million or 0.7%; organic revenue including contingents was $1.237 billion, up 0.7%. The company’s reconciliation excluded $393 million of acquisition-related commissions and fees from the organic revenue calculation. [S4]
Adjusted EBITDAC increased 27.0% to $598 million, while its adjusted margin decreased from 36.7% to 35.7%. Income before taxes rose 23.2%, but its margin fell from 24.2% to 22.9%. These are company-reported measures; adjusted EBITDAC is non-GAAP. [S4]
The organic picture differed between segments. Retail organic revenue grew 1.5% in Q2 (2.5% including contingents). Specialty Distribution organic revenue declined 3.5% (1.6% including contingents), even as reported Specialty Distribution revenue rose 28.1% to $721 million. Management attributed part of Specialty’s organic decline to declining catastrophe-property rates and timing of new business, while also noting other line and investment effects. [S5]
Organic revenue is a company-defined measure that adjusts reported commissions and fees for acquisitions, dispositions, foreign-currency translation and specified litigation effects; the “with contingents” measure adds eligible profit-sharing commissions. Acquisition growth and insurance market conditions can move reported revenue differently from underlying sales activity. The Q2 results do not identify an AI-related problem or a benefit AI could solve. [S4–S5]
Exhibit 1. Selected Q2 2026 company results
USD millions except percentages and per-share amounts; unaudited company release.
| Measure | Q2 2026 | Q2 2025 | Change |
|---|---|---|---|
| Total revenue | $1,676 | $1,285 | +30.4% |
| Organic revenue | $1,176 | $1,184 | −0.7% |
| Organic revenue with contingents | $1,237 | $1,229 | +0.7% |
| Income before taxes | $383 | $311 | +23.2% |
| Income before taxes margin | 22.9% | 24.2% | −1.3 pts |
| Adjusted EBITDAC | $598 | $471 | +27.0% |
| Adjusted EBITDAC margin | 35.7% | 36.7% | −1.0 pt |
Source: [S4], Q2 results and reconciliation tables. Rounded to nearest million; percentages as reported. Organic measures and adjusted EBITDAC are company-defined; adjusted EBITDAC is non-GAAP.
Exhibit 2. Q2 segment comparison
| Segment | Q2 2026 revenue | Reported revenue growth | Organic revenue growth | Organic growth with contingents |
|---|---|---|---|---|
| Retail | Approximately $955 million (inferred) | +35.9% | +1.5% | +2.5% |
| Specialty Distribution | $721 million | +28.1% | −3.5% | −1.6% |
Source: [S5], Q2 2026 segment overview and 10-Q. Rounded segment revenue is inferred from reported total less Specialty Distribution. Organic growth rates are company-reported. Segment growth includes acquisitions; the Specialty results include the effects described in management’s discussion.
What Brown & Brown announced about AI
The July announcement set out four stated aims: accelerate growth, enhance customer experience, improve teammate productivity and strengthen business performance. The company said it would ultimately deploy Claude across about 23,000 teammates and integrate AI into end-to-end workflows in customer service, operations, technology and corporate functions. It also said it would deploy Claude Code across its software engineering organization. [S1]
The company named three external partners for frontier AI, business transformation and governance. It described the VMO as the body that would monitor adoption, measure business impact and ROI, and maintain controls. It did not publicly disclose the VMO’s organization, budget, decision rights, workflow-specific baselines or expansion criteria. [S1]
The September overview described a broader technology journey that included data standardization, platform rationalization and future enterprise-scale enablement for AI-driven growth. It identified innovation and AI, actionable analytics, and production stability and security as technology and data pillars. [S2]
These public materials establish an announced direction, partners and intended governance mechanisms. They do not disclose which business workflow should scale first, what that workflow currently costs, or what outcomes a wider rollout has achieved.
Rival signals
Insurance brokers are beginning to present AI as part of customer-facing expertise. Gallagher announced “Gallagher Blueprint,” which it describes as combining AI analytics, proprietary data and specialist knowledge to create client risk profiles and action plans. It also announced an AI tool for employer-benefits decision-making. These are Gallagher’s product descriptions, not independent evidence of performance. [S6–S7]
The competitor signals make AI-enabled client service strategically relevant. They do not establish customer adoption, improved retention, lower costs, or that any rival has achieved a durable advantage.
Three possible first waves
The following alternatives are constructed for this case from Brown & Brown’s public segment structure and announced aims. They are not publicly announced Brown & Brown projects.
1. Retail service and renewal preparation
An assistant could help local service teams retrieve approved information, summarize a customer’s renewal history, prepare routine communications or assemble a renewal checklist. Local staff would still verify client and policy context and decide what advice to give or send.
This path targets customer experience and staff capacity across a large segment. It also puts AI near client interactions, where an inaccurate or incomplete statement about coverage can harm trust and create errors-and-omissions exposure. The existing public record does not show the volume of these tasks, their baseline response times, or common error and rework rates.
2. Specialty Distribution submission and placement support
An assistant could help specialist teams organize submissions, summarize risks, find relevant internal or market information, or prepare comparisons for human review. Specialists would retain decisions on risk description, insurer approach, coverage and client advice.
This path could support specialized expertise and capacity in a segment with negative Q2 organic growth. However, management attributes some of that decline to market rates and new-business timing. Specialty cases can vary widely; insurer appetite and terms change, and wrong or stale information could undermine the very expertise Brown & Brown sells.
3. Common internal support process
Brown & Brown could start with a routine internal process with defined inputs and outputs—for example, internal knowledge search or recurring administrative support. This may create comparable measures and teach teams how to use common tools before AI reaches more sensitive service workflows.
This path may be easier to evaluate and govern, but could deliver little strategic value if employees do not use the time released for higher-value work. A common process may also fit differently across local businesses and acquired firms.
Your recommendation
Assume you are advising Brown & Brown leadership at the information cutoff. Choose one first scaled workflow or propose a bounded sequence beginning with one of the three alternatives. Explain:
- Why this workflow fits the company’s proposition and the Q2 business context.
- Which tasks AI should assist, which decisions remain with people, and what information may be used.
- How local teams should shape and validate the workflow while the VMO sets shared standards.
- What measures, comparison design and quality thresholds will show value, risk and adoption.
- What evidence would cause leaders to expand, modify, pause or stop the rollout.
Your memo should distinguish cash savings from released capacity and from improvements in customer experience, retention or growth. Do not treat a reported pilot productivity multiplier as a forecast. State any assumptions you make; public sources do not disclose Brown & Brown’s workflow-level costs, volumes or results.
Sources and conventions
S1. Brown & Brown, AI-first transformation announcement, 23 July 2026; AI-first plan, partner roles, VMO, Claude Code pilot reports and forward-looking statement.
S2. Brown & Brown, 2026 Company Overview, Exhibit 99.1, Form 8-K accession 0001193125-26-386867. Accepted by EDGAR 9 September 2026 at 21:17:58 ET; filed 10 September 2026. Used under the 9 September 23:59 ET cutoff. Slides “Decentralized Sales & Service Model,” “Technology & Data,” “Building an AI-powered BRO,” “Our Investment Pillars” and company-scale metrics.
S3. Brown & Brown, FY2025 Form 10-K, accession 0001193125-26-046984, filed 12 February 2026; Business and segment descriptions.
S4. Brown & Brown, Q2 2026 results release, 27 July 2026; results and “Reconciliation of Commissions and Fees to Organic Revenue and Organic Revenue with Contingents.”
S5. Brown & Brown, Q2 2026 Form 10-Q, accession 0001193125-26-318251, filed 27 July 2026, period ended 30 June; MD&A and segment discussion. Also Brown & Brown, Q2 2026 business overview, segment pages.
S6. Arthur J. Gallagher & Co., Gallagher Blueprint announcement, 4 May 2026.
S7. Arthur J. Gallagher & Co., AI benefits decision tool announcement, 14 May 2026.
Company metrics and plans are attributed to company disclosures. All three workflow alternatives and the assignment are instructional constructs. No post-cutoff outcome is included.