Zurich Insurance's AI Strategy: How the $80B Global Carrier Runs Commercial Lines Customer Discovery

TL;DR#

Zurich Insurance Group, the $80B Swiss-headquartered top-five global property and casualty carrier, is rewiring commercial-lines customer discovery around AI-driven conversation rather than broker-mediated forms. Operating in 215 countries and territories with roughly 60,000 employees, Zurich runs one of the most complex specialty and middle-market portfolios in the industry — covering everything from multinational construction programs to cyber, marine, and climate-resilience consulting through Zurich Resilience Solutions (ZRS). Traditional voice-of-customer (VoC) tools built for retail auto and home flatten that complexity into Likert scales, which is why Zurich has publicly committed to generative-AI customer-experience initiatives in its 2023–2025 strategic cycle and its current 2025–2027 "Beyond" plan. The pattern: replace static broker satisfaction surveys with AI-led conversational research that captures risk narrative, climate exposure, and supply-chain vulnerability in customers' own words. For the $1.1 trillion global commercial insurance market — projected by Swiss Re Institute's sigma research to grow ~3.4% real per year through 2027 — Zurich's playbook signals where specialty carriers are headed: conversation as a primary data layer for underwriting, retention, and resilience advisory. This post breaks down how Zurich captures commercial-lines customer narrative across its three operating regions (EMEA, North America, Asia Pacific & LatAm), and what it means for any carrier still relying on forms.

What is commercial insurance AI?#

Commercial insurance AI refers to the use of machine learning, large language models, and conversational agents to underwrite, service, retain, and research commercial policyholders — covering small-business, middle-market, and large-account risk across property, casualty, marine, cyber, and specialty lines. Unlike retail-lines AI (auto, home) where pricing is highly automated and exposures are standardized, commercial AI must handle bespoke risk narratives, broker intermediaries, multi-year program structures, and geographies-specific regulation, which is why conversational research is gaining ground over static survey instruments.

Why specialty commercial lines fail under traditional VoC tools#

Specialty commercial lines fail under traditional voice-of-customer tools because risk is bespoke, intermediated by brokers, distributed across global geographies, and resists the schema-flattening that forms require. A multinational construction program written by Zurich in Zurich, London, and São Paulo cannot be meaningfully evaluated by an NPS question and a five-point dropdown.

Three structural problems break legacy VoC for commercial:

1. Risk complexity defies fixed schemas. A typical Zurich middle-market customer might carry property, general liability, workers' compensation, marine cargo, environmental, cyber, and directors & officers (D&O) coverage simultaneously — often through a captive structure with reinsurance layers. A survey field labeled "How satisfied are you with your policy?" assumes one policy and ignores 80% of the actual customer relationship. Conversational research, by contrast, lets the risk manager talk about the program as a whole, which is the unit of value they actually purchase.

2. The broker layer obscures the customer's voice. In commercial lines, the customer's primary relationship is often with a broker (Marsh, Aon, Willis Towers Watson, Lockton, HUB), not the carrier directly. NPS surveys sent through the broker get filtered, summarized, or dropped. AI-led interviews — which can be deployed to risk managers, CFOs, and corporate insurance buyers directly with broker consent — restore the carrier's line of sight to the actual policyholder. See commercial insurance AI in 2026: a practical guide for brokers, MGAs, and carriers for how the three-sided relationship reshapes the research workflow.

3. Geography fragments the data. Zurich operates in 215 countries and territories — a Munich-based DAX 40 manufacturer, a Singapore-based logistics conglomerate, and a Chicago-based middle-market hospital each face different regulatory regimes, language preferences, and risk profiles. A globally translated survey instrument loses fidelity at every step. Conversational AI handles this natively: it can run multilingual interviews in each customer's working language and reconcile findings into a single global view.

The Lemonade case study on conversational AI in insurance showed what's possible in retail; Zurich is now applying the same pattern at the opposite end of the market.

Inside Zurich's AI customer-discovery program#

Zurich's AI customer-discovery program is structured around three pillars — broker engagement, direct policyholder risk-narrative capture, and large-account strategic research — each backed by conversational agents that replace static surveys with structured dialogue. The architecture mirrors what Zurich CEO Mario Greco has publicly framed as "becoming the customer-led insurer of the future" in the group's 2025–2027 "Beyond" strategic plan.

Broker engagement: replacing the annual broker satisfaction survey#

Broker engagement at Zurich is moving from an annual NPS pulse to a continuous conversational layer that captures broker priorities, friction points, and competitor wins in near-real time. The legacy approach — a once-a-year 30-question Likert-scale survey — typically drew sub-15% completion and arrived too late to act on. According to research published by the Gartner peer insights program on VoC programs broadly, response rates have been declining across categories for the past five years.

A modern broker-engagement program looks more like this: when a broker submits a binder, AI initiates a 5-minute follow-up conversation about why Zurich won (or lost) versus AIG, Chubb, Travelers, or Liberty Mutual; what coverage gaps the broker noticed in the quote; and what the broker's client actually said when reviewing the proposal. Those conversations feed product, underwriting, and distribution teams within 24 hours instead of 12 months. The mechanics mirror the workflow described in Chubb's AI strategy for specialty insurance, where competitive intelligence at the broker level became the highest-leverage research surface.

Direct policyholder risk-narrative capture#

Direct policyholder research at Zurich captures risk narrative from named-insured risk managers, CFOs, and operations leaders — the people who actually live the exposure — through conversational research designed to surface the "why" behind a coverage decision. This is research that a five-point dropdown cannot do.

Three practical applications:

• Pre-renewal risk interviews with the named insured's risk manager, covering changes in operations, M&A activity, climate exposure, cyber posture, and supply-chain shifts. A pre-renewal conversational research outline lets the carrier underwrite the relationship, not just the exposure schedule.
• Post-claim experience interviews that capture what the claims process actually felt like — not just whether NPS moved. These interviews directly feed the voice of customer survey methodology used by claims operations, but in conversation rather than form format.
• Win-loss interviews with prospects who chose a competing carrier, exposing the real reasons (price was rarely the only one) versus the broker's filtered version. Run them via the win-loss interview template for repeatable structure.

Large-account strategic research#

Large-account research at Zurich — for Fortune 500 customers in its International Programs business — is where conversational AI delivers the most dramatic uplift over surveys, because these accounts are too valuable, too complex, and too few in number for traditional sampling. A typical multinational program touches risk managers in 30+ countries; sending all of them a survey produces noise. Conducting structured AI-led interviews with each, then synthesizing the cross-geography view, produces the actual signal.

This is the workflow Zurich's underwriting and distribution teams have publicly described as core to the "customer-led" pivot, and it parallels what carriers like Liberty Mutual's AI customer experience modernization and Travelers' AI risk modeling and conversational underwriting have built. See also the parallel approach inside AIG's AI commercial insurance conversational underwriting and Allianz's AI customer research at Europe's largest insurer for how peer European-headquartered carriers are running the same play.

Zurich Resilience Solutions: climate, supply chain, and cyber discovery via conversation#

Zurich Resilience Solutions (ZRS) is the carrier's risk-engineering and advisory arm — and it's where AI-led conversational research delivers the clearest commercial return because resilience advisory is fundamentally a discovery business, not an indemnity business. ZRS is positioned by the group as a growth engine in the 2025–2027 plan, with the explicit goal of moving Zurich from indemnifier to advisor on climate, supply chain, and cyber resilience.

Climate resilience: capturing the exposure story#

Climate resilience research at ZRS goes beyond catastrophe-model outputs to capture how customers actually experience and adapt to physical climate risk on the ground. CAT models tell Zurich the probabilistic loss; conversation tells Zurich what the customer is doing about it.

A typical conversational research program with a manufacturing customer might surface: which facilities have flood-mitigation capex queued for 2026; which suppliers in Southeast Asia are at heat-stress risk; which contracts have force-majeure language that won't hold up under a 1-in-50-year event. None of that fits a checkbox. According to a McKinsey climate risk and response report, corporates that have integrated physical-climate exposure into operational planning capture roughly 4–6x the resilience ROI of those that haven't — but identifying who's in which group requires structured conversation, not a survey.

Supply chain resilience: from BCP forms to BCP conversations#

Supply chain resilience research replaces the annual Business Continuity Planning (BCP) questionnaire with structured interviews that capture tier-2 and tier-3 supplier dependencies, single-points-of-failure, and contractual buffer stock arrangements. The BCP form is an industry punchline: every risk manager fills it out, almost no one acts on it. Conversation surfaces what the form omits.

This pattern is identical to what we've outlined in AI customer communications in the insurance industry: 2026 state of the industry report, where the structural failure of forms in commercial intake was quantified.

Cyber resilience: surfacing the controls story#

Cyber resilience research captures the narrative around security posture — patch cadence, MFA coverage, third-party access, incident-response maturity — in a way that static cyber questionnaires cannot, because cyber risk is fundamentally a story about controls and culture, not a checklist. According to IBM's Cost of a Data Breach Report, the average breach cost crossed $4.88M in 2024; carriers that can capture cyber-posture narrative at bind and renewal can both price more accurately and offer practical hardening advice. ZRS's cyber consulting model treats every renewal as a fresh discovery conversation, not a copy-paste of last year's questionnaire — a workflow paralleling what we've described in AI insurance fraud detection in 2026.

What this signals for the $1T global commercial insurance market#

Zurich's pivot signals that the entire $1T+ global commercial insurance market is moving from form-driven, broker-mediated VoC toward AI-led conversational customer discovery as a core competitive surface — and carriers that don't follow will lose ground on retention, cross-sell, and resilience advisory revenue.

Three implications for the rest of the market:

1. The "named-carrier conversational case study" is becoming a category. From Hartford's AI strategy for small business with conversation to Chubb's $260B specialty playbook to American Family's top-10 carrier modernization to Prudential's life-insurance conversational policyholder research, every major carrier is now publicly committed to a conversational customer-research layer. The carriers that don't talk about it publicly are doing it quietly.
2. MGAs and program businesses will go first on conversational underwriting. Smaller specialty MGAs have less legacy survey infrastructure to dismantle, which is why programs like Pie Insurance's AI-first workers' comp underwriting, Next Insurance's AI-first SMB playbook, and Cover Genius's embedded insurance AI strategy are showing what's possible at sub-billion-dollar scale before the global carriers fully roll it out.
3. The broker's role shifts from gatekeeper to co-pilot. When carriers run direct AI-led interviews with the named insured (with the broker's consent and visibility), the broker stops being the bottleneck on customer signal — and starts being the strategic interpreter of that signal. The brokers who lean into this win; the ones who guard the customer relationship lose share.

For carriers and brokers operating in this space, the practical starting point is a single high-leverage research surface — typically renewal-risk interviews or broker win-loss — run via an AI interviewer agent built for customer research at scale. Many programs start with an insurance quote conversation flow and an insurance coverage explainer agent for the front-of-funnel customer experience, then layer in renewal and post-claim conversations once the workflow is proven. Teams running this in production are built for CX teams and product teams at carriers, MGAs, and insurance distribution platforms.

Frequently Asked Questions#

What is Zurich Insurance Group's AI strategy in commercial lines?#

Zurich Insurance Group's AI strategy in commercial lines centers on replacing static, broker-mediated forms and surveys with AI-led conversational research and underwriting workflows across its broker engagement, direct policyholder, and large-account research programs. The strategy is part of the 2025–2027 "Beyond" plan under CEO Mario Greco and integrates with Zurich Resilience Solutions for climate, supply chain, and cyber advisory. The practical effect: faster signal, deeper risk narrative, and richer cross-sell intelligence than legacy VoC can produce.

How big is the global commercial insurance market and how fast is it growing?#

The global commercial insurance market is roughly $1.1 trillion in annual gross written premium and is projected to grow approximately 3.4% in real terms per year through 2027, according to Swiss Re Institute sigma research. Specialty lines (cyber, climate, marine, financial lines) are outgrowing the average, which is the segment where Zurich, Chubb, AIG, and Allianz compete most directly. AI-led customer discovery is concentrated in these higher-margin specialty segments because the per-account economics support it.

Why don't traditional surveys work for commercial insurance customers?#

Traditional surveys don't work for commercial insurance customers because risk is bespoke, programs span multiple lines and geographies, brokers sit between the carrier and the named insured, and the highest-value information ("we just acquired a plant in Vietnam and we're not sure how to handle the marine exposure") is fundamentally a narrative, not a field. Five-point Likert scales can't capture this, which is why carriers like Zurich, AIG, and Allianz have moved toward conversational research that lets risk managers, CFOs, and operations leaders speak in their own words.

What is Zurich Resilience Solutions (ZRS)?#

Zurich Resilience Solutions (ZRS) is Zurich Insurance Group's risk-engineering and advisory arm, focused on climate, supply chain, cyber, and operational resilience consulting for commercial customers. ZRS is a growth pillar in Zurich's 2025–2027 strategic plan and uses conversational research and AI-led discovery to move the carrier's relationship from indemnifier to strategic risk advisor. The advisory model is structurally well-suited to conversational AI because resilience is a discovery business — you cannot indemnify what you have not yet articulated.

How does conversational AI improve broker satisfaction research?#

Conversational AI improves broker satisfaction research by replacing low-completion annual NPS surveys with structured, on-demand interviews triggered by binder events, quote rejections, or competitive losses. Instead of a 30-question Likert form that arrives 11 months too late, the broker has a 5-minute conversation when the signal is fresh — explaining what the client said, what coverage gap surfaced, and which competing carrier (AIG, Chubb, Travelers, Liberty Mutual, Allianz) was the runner-up. The output feeds product, underwriting, and distribution teams within 24 hours.

Which other global carriers are running similar AI customer-research programs?#

Other global carriers running similar AI customer-research programs include AIG, Allianz, Chubb, Travelers, Liberty Mutual, The Hartford, and Prudential — each at different stages of public commitment but all moving in the same direction. Specialty MGAs and AI-native carriers (Lemonade, Hippo, Root, Pie, Next, Branch, Cover Genius) are typically further along on conversational customer experience because they have less legacy infrastructure to dismantle. The competitive question is no longer whether to do this — it's how fast.

Conclusion#

Zurich Insurance Group's pivot to AI-led conversational customer discovery in commercial lines is one of the clearest signals yet that the era of broker-mediated, form-based voice-of-customer is ending in commercial insurance AI. An $80B Swiss-headquartered carrier operating in 215 countries does not casually rewrite its customer-research stack — and the fact that Zurich, alongside AIG, Allianz, Chubb, Travelers, and Liberty Mutual, is publicly building toward conversational research and underwriting tells you the direction of the entire $1.1 trillion global commercial insurance market.

The carriers, MGAs, and brokers who win the next decade will be the ones who can capture commercial-line risk narrative at the speed of conversation — across broker engagement, named-insured renewal research, large-account strategic discovery, and ZRS-style resilience advisory. That requires more than buying an AI vendor; it requires re-architecting the research workflow around dialogue, not dropdowns.

Perspective AI was built for exactly this: an AI interviewer that can run hundreds of structured commercial-lines customer conversations simultaneously, in the language and time zone the customer prefers, with automatic synthesis into the actionable signal product, underwriting, and distribution teams need. Start a research study, see what teams are running today in the studies gallery, or review pricing to scope a pilot — most carriers begin with one of broker win-loss, pre-renewal risk interviews, or post-claim NPS-replacement, and expand from there.