量身 AI← 案例库← Case library
案例 · AI 落地方案CASE · AI DEPLOYMENT PLAN · 挑战者银行 · 多伦多Challenger bank · Toronto

恒信银行 Equitable Bank / EQ BankEQB (Equitable Bank / EQ Bank)

从员工聊天机器人,走到更快的放款From a staff chatbot to faster lending
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 已证明的收益在员工支持;对房贷与商业贷款审批的影响未披露。The proven gain is in staff support; the effect on mortgage and commercial approvals is not disclosed.
起步顺序Order of work 先把文件数据理干净,再加速房贷审核,最后扩展到商业贷款与反欺诈。Make the document data clean first, then speed up mortgage review, then extend to commercial lending and fraud.

公开事实Public facts 已核对Checked

  • 1,947
    2025 年员工数Employees in 2025规模小、决策快的银行A small, quick-moving bank
  • C$53.49B
    2025 年总资产Total assets in 2025按资产为加拿大第七大一级特许银行Seventh-largest Schedule I bank by assets
  • ~4M
    服务的加拿大人Canadians served房贷、商业贷款、数字存款Mortgages, commercial lending, digital deposits
  • C$354M
    2025 年净利润Net income in 2025营收 12.6 亿加元Revenue C$1.26B

对员工已证明,对放款未证明Proven for staff, unproven for lending 判断Judgment

  • 审批速度Approval speed:审批周转时间,目标由 EQB 自定Approval turnaround, a target for EQB to set
  • 对房贷与商业贷款审批的影响:未披露Effect on mortgage and commercial approvals: not disclosed
  • 员工侧的收益是真的;放款才是银行赚钱的地方。The gain is real for staff; lending is where the bank earns.

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

没有研究院,就买成熟工具;只自建数据与流程。With no research lab, buy mature tools; own only data and process.

Thomas Davenport

整家企业怎么变Whole-enterprise change

房贷审核是要改的流程;点名业务负责人。Mortgage review is the process to change; name its business owner.

Erik Brynjolfsson

价值与度量Value & measurement

先定审批时长基线;投入培训让员工侧收益传到放款。Baseline approval time; fund training so staff gains reach lending.

Ethan Mollick

人怎么用How people use it

用杂乱的估值与收入文件测出它的失败点。Test on messy appraisals and income documents to find where it fails.

Ziad Obermeyer

算法在优化什么What it optimizes

审批:写清真实目标与代理指标;按人群检查结果。For approvals, name the real goal and proxy; check results by group.

Andrew Ng

节奏与方法Pace & method

选一个三个月内出数字的房贷文件任务。Pick one mortgage document task that shows numbers in three months.

保留的分歧Disagreement kept · 小而快 vs 决策审查Small and fast vs. decision review

Andrew Ng:三个月内上线一个文件任务。Ship one document task in three months.

Ziad Obermeyer:审批模型先过目标检查。Approval models pass a target check first.

→ 文件任务走快速通道;审批决定完整审查。Document tasks go fast; approval decisions get full review.

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

给审批速度定基线Baseline approval speed

量出房贷与商业贷款当前的周转时间。Measure today's turnaround on mortgages and commercial loans.

2

加速文件审核Speed up document review

抽取并核对收入与估值文件;由人决定。Extract and check income and appraisal documents; people decide.

3

扩展到商业贷款与反欺诈Extend to commercial and fraud

在新 CRM 之上扩展,再加身份验证与反欺诈。Build on the new CRM, then add identity and fraud checks.

场景组合Scenarios 建议Recommendation

场景Scenario做法Approach状态Status
房贷文件审核Mortgage document review
提效Efficiency体验Experience
抽取并核对收入与估值文件;由核保人决定。Extract and check income and appraisal files; underwriters decide.建议Proposed
商业贷款全生命周期Commercial loan lifecycle
提效Efficiency降风险Risk
在新的 Dynamics 365 CRM 上做文件抽取与风险预警。Document extraction and risk alerts on the new Dynamics 365 CRM.建议 · 基于新 CRMProposed · builds on the new CRM
EQ Bank 开户与反欺诈EQ Bank onboarding and fraud
降风险Risk体验Experience
为无网点银行强化线上身份验证与欺诈筛查。Strengthen online identity checks and fraud screening for a branchless bank.建议 · 完整审查Proposed · full review
统一 AI 赋能平台One AI enablement platform
提效Efficiency
由一位负责人统管微软与谷歌工具,追踪使用与价值。One owner for the Microsoft and Google tools, with usage and value tracked.建议 · 公开招聘指向此方向Proposed · public hiring points this way

12 个月路线12-month roadmap 建议Recommendation

数据Data

第 1–4 月Months 1–4
  • 文件数据整理Document data cleaned
  • 审批基线Approval baselines
  • 模型清单与分级Model inventory, tiers
◆ 基线确认Baselines agreed

试点Pilot

第 5–7 月Months 5–7
  • 房贷审核任务上线Mortgage review task live
  • 对照组Control group
  • 边界案例记录Edge-case log
◆ 指标达成Targets met

上线Go live

第 8–10 月Months 8–10
  • 推广到全部房贷团队Rollout to all mortgage teams
  • 审批由人决定Person decides approvals
  • 上线即监控Monitoring from day one
◆ 风险合规签字Risk sign-off

扩展Extend

第 11–12 月Months 11–12
  • 商业贷款试点Commercial loan pilot
  • 反欺诈范围确定Fraud checks scoped
  • 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

  • 模型清单对齐 OSFI E-23(2027-05-01 生效)Model inventory aligned to OSFI E-23 (effective 2027-05-01)
  • 审批与反欺诈模型:完整审查Approval and fraud models: full review
  • 审批决定由人作出A person decides approvals

三层指标Three-layer metrics 建议Recommendation

采用Adoption每周使用工具的核保人Underwriters using the tool weekly · 工具处理的文件数Documents processed by the tool
过程Process文件审核时长Document review time · 人工修正率Human correction rate · 审批周转时间Approval turnaround
结果Outcome人均放款笔数Loans closed per underwriter · 单笔放款成本Cost per funded loan

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

风险Risks 判断Judgment

收益止步于员工支持Gains that stop at staff support

聊天机器人节省的工作量,若不改放款流程,传不到放款。Chatbot savings do not lift lending unless the loan process changes.

文件杂乱Messy documents

估值与收入文件五花八门;上线前用真实文件测试。Appraisals and income files vary; test on real ones before rollout.

审批中的代理指标Proxy goals in approvals

模型可能偏向好办的文件,而非好贷款;两个目标都要写清。A model may favour easy files, not good loans; name both goals.

前 90 天First 90 days 建议Recommendation

第 1–2 周Weeks 1–2点名放款负责人;开始定基线Name the lending owner; start baselines
第 3–4 周Weeks 3–4盘点工具与在用模型Inventory tools and live models
第 5–10 周Weeks 5–10整理样本文件;选定一个任务Clean sample documents; pick one task
第 13 周Week 13首次复盘First review
今天只需要一个决定:One decision today: 为房贷审批速度指定一位负责人name one owner for mortgage approval speed

我们可能错在:We could be wrong if: 审批时长其实受信贷政策而非文书限制,文件处理更快也改变不大。approval time is already limited by credit policy rather than paperwork, so faster documents change little.

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]