量身 AI← 案例库← Case library
案例 · AI 落地方案CASE · AI DEPLOYMENT PLAN · 人寿与健康保险 · 多伦多Life and health insurance · Toronto

加拿大永明人寿 Canada LifeCanada Life

从「会回答」的助手,走到「能办完」的助手From an assistant that answers to one that completes
基于公开资料的分析,与该企业没有合作关系。Based on public information; no affiliation with the company.

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 ↗

一句话判断The verdict CaLi 会回答、会查状态,但还不能办完一笔业务;AI 培训证据也偏薄。Cali answers questions and looks up status; it does not yet finish a transaction, and training evidence is thin.
起步顺序Order of work 先补齐培训与治理证据,再做一个简单业务办结,最后进入团体福利。Fix training and governance evidence first, then one simple transaction, then group benefits.

公开事实Public facts 已核对Checked

  • 186,000+
    CaLi 截至 2026 年 10 月处理的对话Conversations handled by CaLi by October 2026约 17 万次与计划成员、1.6 万次与顾问About 170,000 with members, 16,000 with advisors
  • 83%
    客服中心聊天机器人问题解决率Chatbot problem-resolution rate in the contact centre客服中心 AI 在 2025 上半年节省约 750 万加元Contact-centre A I saved about C$7.5M in H1 2025
  • 27 of 30
    Evident 2026 年保险 AI 指数排名Rank on Evident's 2026 insurance AI index加拿大主要寿险公司中最低Lowest among major Canadian life insurers
  • US$150M
    集团 2026 年对 Sagard AI 基金的承诺Group commitment to the Sagard AI Fund, 2026与 Power Corp、IGM 联合With Power Corp. and IGM

CaLi 的四级阶梯Four rungs for CaLi 判断Judgment

  • 回答Answer:成员与顾问提问,CaLi 作答。已上线。Members and advisors ask; CaLi replies. Live today.
  • 查询Look up:个人保险申请的实时状态。2026 年 10 月新增。Live status of an individual insurance application. Added October 2026.
  • 办完Complete:一项简单请求端到端办完,由人确认。建议。One simple request end to end, a person confirming. Proposed.
  • 跑流程Run a process:团体福利理赔与咨询,由业务条线负责。Group benefits claims and enquiries, owned by the business line.

专家组裁决Expert panel 依据公开观点的推演Inference from public views

Foteini Agrafioti

技术与业务战略对齐Strategy alignment

点名一位 CaLi 路线图负责人,覆盖集团、加拿大永明与业务条线;每一级都挂目标。Name one owner of CaLi's roadmap; tie each rung to a business target.

Erik Brynjolfsson

价值与度量Value & measurement

给 18.6 万次对话定基线:哪些替代了电话,哪些创造了新需求?Baseline the 186,000 conversations: calls replaced versus demand created.

Ziad Obermeyer

算法在优化什么What it optimizes

CaLi 碰理赔和受益人之前,先写真实目标与代理指标;电话少不等于结果好。Before claims or beneficiaries, write the real goal and the proxy.

Ethan Mollick

人怎么用How people use it

在真实受益人与理赔请求上测出 CaLi 的失败点;边界摸清前由人确认。Test where CaLi fails on real requests; a person confirms until mapped.

Andrew Ng

节奏与方法Pace & method

选一种请求,三到六个月内出数字,并写明下一步复制到哪。Pick one request type that shows numbers in three to six months.

Thomas Davenport

整家企业怎么变Whole-enterprise change

排名垫底是企业层面的信号:培训、治理与流程,而不是某个工具。A last-place rank is an enterprise signal, not a tool problem.

保留的分歧Disagreement kept · 先赢一单还是先补底Win fast vs. fix the base

Andrew Ng:先快速办成一个简单业务,再复制。Win one simple transaction fast, then copy.

Thomas Davenport:排名垫底是企业问题;先补培训与治理。Fix training and governance first.

→ 两者并行:单业务试点与治理基线同时启动。Run both together: one pilot and a governance baseline.

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

1

补上治理与培训缺口Close the governance gap

建立 AI 培训记录与模型清单,让进步在外部评估中看得见。AI training evidence and a model inventory, so progress shows in outside reviews.

2

让 CaLi 办完一件事Let CaLi finish one request

先从受益人变更开始,每个结果由人确认。Start with a beneficiary change; a person confirms every outcome.

3

延伸到团体福利Carry it into group benefits

理赔与咨询量大;审查通过后再延伸。Heavy claims and enquiry volume; extend only after review.

场景组合Scenarios 建议Recommendation

场景Scenario做法Approach状态Status
CaLi 办理受益人变更Beneficiary changes in CaLi
体验Experience提效Efficiency
让 CaLi 从答问升级为办完受益人变更,由人确认。Move CaLi from answering to completing a beneficiary change, a person confirming.建议 · 基于已上线 CaLiProposed · builds on live CaLi
团体福利理赔受理Group benefits claims intake
提效Efficiency体验Experience
为计划成员准备并提交理赔,由理赔员逐件核对。Prepare and submit plan-member claims; an adjudicator checks each one.建议Proposed
顾问申请助手Advisor application assistant
提效Efficiency增收Revenue
顾问已可获得流程答复;延伸到准备申请材料。Advisors already get process answers; extend to preparing applications.已上线(公开)· 延伸为建议Live (public) · extension proposed
培训与治理计划Training and governance programme
降风险Risk
员工 AI 培训与模型清单,让进步体现在外部排名中。Staff A I training and a model inventory, so progress shows in outside rankings.建议Proposed

12 个月路线12-month roadmap 建议Recommendation

奠基Foundation

第 1–3 月Months 1–3
  • CaLi 对话基线Baseline CaLi conversations
  • 模型清单与风险分级Model inventory and risk tiers
  • 培训计划初稿Training plan drafted
◆ 基线与分级定稿Baselines and tiers final

试点Pilot

第 4–6 月Months 4–6
  • 受益人变更,影子模式Beneficiary change, shadow mode
  • 对照组测量Control group measured
  • 失败边界测试记录Failure-edge test log
◆ 边界摸清Edge mapped

上线Go live

第 7–9 月Months 7–9
  • 受益人变更上线Beneficiary change live
  • 关键步骤人工确认Person confirms key steps
  • 上线即监控Monitoring from day one
◆ 风险合规签字Risk sign-off

扩展Extend

第 10–12 月Months 10–12
  • 团体福利准入评审Group benefits gate review
  • 员工培训推广Staff training roll-out
  • 年度重排优先级Annual re-ranking
◆ 进入下一年Next year begins

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
  • 涉及福利或资金的步骤由人确认A person confirms any step that moves benefits or money
  • 按客户群分别检查结果,而非只看总体Results checked by customer group, not only overall

三层指标Three-layer metrics 建议Recommendation

采用Adoption成员与顾问周活跃使用Weekly active use, members and advisors · 在 CaLi 发起的请求数Requests started in CaLi
过程Process按请求类型的解决率Resolution rate by request type · 处理时长Handling time · 人工确认通过率Human confirmation pass rate
结果Outcome对照基线的减少来电Calls avoided against baseline · 单笔办结成本Cost per completed request

目标值由企业自定。Targets are the company's to set.

风险Risks 判断Judgment

把少接的电话算成节省Counting avoided calls as savings

没有对照组,减少的来电与新增的需求会混在一起。Without a control group, calls avoided and demand created get mixed up.

CaLi 越过能力边界Cali acts beyond its edge

办事与答问的出错方式不同;先在真实请求上测试。Completing tasks fails differently from answering; test on real requests first.

排名变成头条A ranking becomes a headline

垫底排名需要看得见的培训与治理进展。A last-place index rank needs visible training and governance progress.

前 90 天First 90 days 建议Recommendation

第 1–2 周Weeks 1–2点名 CaLi 路线图负责人;定基线Name the CaLi roadmap owner; fix baselines
第 3–4 周Weeks 3–4盘点模型与培训证据Inventory models and training evidence
第 5–10 周Weeks 5–10风险分级;快速通道;选定一种请求Risk tiers; fast lane; pick one request type
第 13 周Week 13首次复盘First review
今天只需要一个决定:One decision today: 指定 CaLi 从答问到办结的路线图负责人name the owner of Cali's roadmap from answers to completed requests

我们可能错在:We could be wrong if: 客户即使在 CaLi 能办时仍打电话办涉及资金的变更,或排名反映的是披露而非实际做法。customers still phone for money-related changes even when CaLi can do them, or the rank reflects disclosure rather than practice.

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.

Telegram @johnsunlei邮件 Email [email protected]