英文旁白、无字幕的 3–4 分钟管理层简报。3–4 minute executive brief, English narration, no captions. 在 YouTube 打开Open on YouTube ↗
起步顺序Order of work 先用加拿大基线重建节省测算,再复制理赔,随后核保,最后经纪工具。Rebuild the savings case with Canadian baselines, then copy claims, then underwriting, then broker tools.
公开事实Public facts 已核对Checked
- ~£100M英国车险理赔转型节省Saved by U.K. motor claims transformation接近此数;集团层面结果Nearly; group-level result
- 150+为理赔训练的机器学习模型Machine-learning models trained for claims集团层面,机器学习已用超 10 年Group level, built over 10+ years of machine learning
- Half生成式 AI 核保工具下的医疗报告审查时间Medical-report review time, with a gen-AI underwriting tool2025 年 11 月先用于个人寿险;2026 年扩展到重疾险Personal life from Nov 2025; critical illness in 2026
- £408MAviva 加拿大 2025 年运营利润Aviva Canada operating profit, 2025同比增长 49%Up 49% year on year
三条要复制的流程,每条都有人把关Three flows to copy, each with a person 判断Judgment
- 理赔Claims:简单理赔进入Simple claim arrives → 智能体整理并提议Agent assembles and proposes → 理赔员确认Adjuster confirms
- 核保Underwriting:报告与文件进入Reports and documents arrive → 助手摘要并标注Assistant summarises and flags → 核保员决定Underwriter decides
- 经纪报价Broker quoting:经纪人请求报价Broker asks for a quote → 助手预填风险信息Assistant pre-fills the risk → 核保员批准Underwriter approves
专家组裁决Expert panel 依据公开观点的推演Inference from public views
Foteini Agrafioti
集团技术与加拿大业务共用一页纸:谁选场景,谁对结果负责。One page: who picks use cases, who owns results in Canada.
Erik Brynjolfsson
英国的节省不会原样搬到加拿大;先定加拿大基线和对照组。U.K. savings will not transfer one for one; set Canadian baselines.
Ziad Obermeyer
定价与反欺诈决定谁付多少:写清真实目标与代理指标,并按客户群检查。For pricing and fraud, name the real goal and the proxy.
Ethan Mollick
在真实的加拿大简单理赔上测试,先知道它在哪里失败再放大。Test the claims agent on real simple claims before widening.
Andrew Ng
一种简单理赔六个月内上线,有点名的理赔负责人,并写明下一步复制到哪。One simple-claim type live in six months, with a named lead.
Thomas Davenport
这是理赔与核保流程的改变,且经由经纪渠道;点名流程负责人,而非 IT 负责人。A process change in claims and underwriting; name process owners.
Andrew Ng:快速复制英国已验证的试点;势头重要。Copy the proven U.K. pilot quickly.
Ziad Obermeyer:加拿大的定价与反欺诈复用前须重新检查目标与公平性。Re-check Canadian pricing and fraud first.
→ 理赔复制以影子模式先启动;定价与反欺诈等准入关。Claims start in shadow mode; pricing and fraud wait for the gate.
六位专家是谁、怎样思考、有什么铁律 →Who the six experts are, how they think, their rules →
三个目标Three priorities 建议Recommendation
在加拿大验证英国做法Prove the U.K. playbook in Canada
用加拿大基线与对照组重建节省测算。Rebuild the savings case with Canadian baselines and a control group.
让智能体处理简单理赔Put an agent on simple claims
先影子模式;每个结果由理赔员确认。Shadow mode first; an adjuster confirms every outcome.
给经纪人一个助手Give brokers an assistant
报价与核保支持;按我们的判断,经纪渠道承担多数分销。Quote and underwriting support, since brokers carry most distribution, in our reading.
场景组合Scenarios 建议Recommendation
| 场景Scenario | 做法Approach | 状态Status |
|---|---|---|
| 简单理赔智能体Simple-claims agent 提效Efficiency体验Experience | 整理并提议简单理赔结果,由理赔员确认。Assemble and propose simple claim outcomes; an adjuster confirms. | 建议 · 借鉴英国智能体测试Proposed · adapts the U.K. agent tests |
| 核保文件助手Underwriter document assistant 提效Efficiency | 把集团将报告审查时间减半的工具改造到加拿大险种。Adapt the group's tool that halves report review time to Canadian lines. | 建议Proposed |
| 车险定价与反欺诈Motor pricing and fraud 降风险Risk增收Revenue | 用定价与反欺诈检查守住 94.7% 的个人险综合成本率。Defend the 94.7% personal-lines combined ratio with pricing and fraud checks. | 建议 · 完整审查Proposed · full review |
| 经纪报价助手Broker quoting assistant 提效Efficiency增收Revenue | 预填风险信息并标出缺漏,让经纪人更快报价。Pre-fill risk details and flag gaps so brokers quote faster. | 建议Proposed |
12 个月路线12-month roadmap 建议Recommendation
奠基Foundation
- 加拿大基线与对照组Canadian baselines and control
- 模型清单与风险分级Model inventory and risk tiers
- 一页纸:集团与加拿大One page: group and Canada
试点Pilot
- 简单理赔影子模式Simple claims in shadow mode
- 与理赔员结果对比Adjuster comparison measured
- 失败边界测试记录Failure-edge test log
上线Go live
- 理赔智能体上线,人工确认Claims agent live, person confirms
- 经纪助手试点Broker assistant pilot
- 上线即监控Monitoring from day one
扩展Extend
- 定价与反欺诈准入评审Pricing and fraud gate review
- 跨险种复用Reuse across lines
- 年度重排优先级Annual re-ranking
治理与合规Governance 建议Recommendation
- 模型清单;与法律顾问确认 OSFI E-23 适用性Model inventory; confirm OSFI E-23 applicability with counsel
- 定价与反欺诈模型:完整审查,含按客户群检查Pricing and fraud models: full review, including by customer group
- 影响理赔结果的步骤由人确认A person confirms outcomes that affect a claim
三层指标Three-layer metrics 建议Recommendation
| 采用Adoption | 各团队周活跃使用Weekly active use by team · 智能体提议的理赔数Claims proposed by the agent |
|---|---|
| 过程Process | 智能体提议被采纳率Agent proposal accepted rate · 处理时长Handling time · 经纪报价周期Broker quote turnaround |
| 结果Outcome | 对照加拿大基线的节省Savings against Canadian baseline · 单件理赔成本Cost per claim |
目标值由企业自定。Targets are the company's to set.
风险Risks 判断Judgment
假定英国节省可直接搬来Assuming U.K. savings transfer
理赔结构、经纪渠道与规则都不同;在本地重建测算。Different claims mix, brokers and rules; rebuild the case locally.
定价优化的是代理指标Pricing optimises the proxy
模型可能追逐比率而非公平结果;上线前写清两者。A model can chase a ratio, not fair outcomes; name both before launch.
经纪渠道阻力Broker friction
经纪人不用的工具没有价值;与少数经纪人共同设计。A tool brokers will not use creates no value; co-design with a few brokers.
前 90 天First 90 days 建议Recommendation
| 第 1–2 周Weeks 1–2 | 点名加拿大基线负责人Name the owner of Canadian baselines |
|---|---|
| 第 3–4 周Weeks 3–4 | 盘点所有在用模型与工具Inventory every live model and tool |
| 第 5–10 周Weeks 5–10 | 风险分级初稿;选定一种理赔Risk tiers drafted; pick one claim type |
| 第 13 周Week 13 | 首次复盘First review |
我们可能错在:We could be wrong if: 加拿大的理赔与经纪渠道差异大到英国智能体需要重做,或加拿大基线显示收益早已取得。Canadian claims and brokers differ enough that the U.K. agent needs a rebuild, or Canadian baselines show the gains were already captured.
出处Sources
本页是基于公开资料的分析,与所述企业没有合作关系,不代表其官方立场或内部计划;企业数字多来自其自述或媒体转述;「专家组」是依据六位专家公开观点的推演,不是他们的原话,也不代表他们或其雇主的立场。资料整理日期:2026-11-14。This page is an analysis based on public information. It has no affiliation with the company and does not represent its official position or internal plans. Company figures come mostly from its own statements or media reports. The expert panel is an inference from six experts' public views, not their words, and does not represent them or their employers. Compiled 2026-11-14.
想给你的公司也做一份?Want one for your company?
同样的方法:公开事实、专属判断、12 个月路线和前 90 天。公司大小都可以聊,中文沟通。Same method: public facts, a company-specific judgment, a 12-month roadmap and the first 90 days. Any size of business; we can talk in Chinese or English.