量身 AI← 案例库← Case library
案例 · AI 落地方案CASE · AI DEPLOYMENT PLAN · 新闻与数字媒体 · 多伦多News and digital media · Toronto

Torstar · 多伦多星报Torstar · Toronto Star

从记者自建的 AI 工具,到全编辑室的工作流From a reporter's AI tool to a newsroom-wide workflow
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.
视频制作中,上线后会出现在这里。Video coming soon — it will appear here.
一句话判断The verdict AI 只在少数记者手里管用;全编辑室还没有共用的工作流。AI works for a few reporters; there is no shared workflow for the whole newsroom.
起步顺序Order of work 先把调查工具变成共用工具,再连接订阅,最后准备内容授权立场。Turn the investigation tool into a shared tool first, then link it to subscriptions, then prepare a licensing position.

公开事实Public facts 已核对Checked

  • 4,242
    用 AI 辅助分析的市政文件Municipal documents analysed with AI assistance用于「Strong Mayors」调查For the Strong Mayors investigation
  • 216
    该调查覆盖的市镇Municipalities covered by that investigation该流程正沉淀为可复用工具The process is being turned into a reusable tool
  • 2,500+
    员工人数(公司自述)Employees, as the company states私有化;无公开财报Private; no public financials
  • $558M
    第三方估算营收Estimated revenue, third-party estimate未经财报核实Not verified against filings

从一次调查到全编辑室工具From one investigation to a newsroom tool 判断Judgment

  • 现在Today:4,242 份市政文件4,242 council documents → AI 辅助分析AI-assisted analysis → 一篇调查One investigation
  • 建议Proposed:任何信息公开文件或大数据集Any FOI file or large dataset → 共用文档工具Shared document tool → 编辑核验后发稿Editor checks, story ships

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

把工具挂到编辑部与订阅目标;编辑与技术的权责写成一页纸。Tie the tool to newsroom and subscription targets; one page of ownership across editors and tech.

Erik Brynjolfsson

价值与度量Value & measurement

按每篇调查工时设基线并做对比;增强记者,预算含培训。Baseline hours per investigation with a comparison; augment reporters, and fund training too.

Ziad Obermeyer

算法在优化什么What it optimizes

AI 决定推荐什么时,写清真实目标与代理指标;所有工具建清单。Where AI picks what to recommend, name the real goal and the proxy; inventory every tool.

Ethan Mollick

人怎么用How people use it

记者已在试用;给他们安全空间,并在真实文件上测出工具失败点。Reporters already experiment; give them a safe space, and test where the tool fails on real files.

Andrew Ng

节奏与方法Pace & method

选一个版面、一类文件;三到六个月内给出节省工时,再复制。Pick one desk and one file type; show hours saved in three to six months, then copy.

Thomas Davenport

整家企业怎么变Whole-enterprise change

改变的是做稿的流程,不只是工具;点名一位编辑为流程负责人。This changes how stories are made, not just a tool; name an editor as the process owner.

保留的分歧Disagreement kept · 节奏与控制Pace vs. control

Ethan Mollick:让记者去试。Let reporters experiment.

Ziad Obermeyer:扩散前先建清单。Inventory tools before spread.

→ 低风险实验现在开放;关键稿件用途须审查。Low-stakes experiments open now; story-critical uses reviewed.

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

把工具变成共用Make the tool shared

把调查流程沉淀为各版面可用的文档分析工具。Turn the investigation workflow into a reusable document tool for every desk.

2

连接订阅业务Link it to subscriptions

个性化推荐与流失预警,对照基线衡量。Personalised recommendations and churn alerts, measured against a baseline.

3

准备授权立场Prepare a licensing position

在下一轮谈判前,弄清内容对 AI 公司的价值。Know what content is worth to A I companies before the next negotiation.

场景组合Scenarios 建议Recommendation

场景Scenario做法Approach状态Status
编辑室文档工具Newsroom document tool
提效Efficiency体验Experience
把 Strong Mayors 的流程做成工具,供各版面处理信息公开文件与大数据集。Turn the Strong Mayors workflow into a tool for FOI files and large datasets, for every desk.建议 · 基于已有流程Proposed · builds on a live workflow
订阅流失预警Subscriber churn alerts
增收Revenue
识别可能退订的读者,向其推荐更匹配的内容。Flag readers likely to cancel and offer them better-matched stories.建议Proposed
授权谈判准备Licensing readiness
增收Revenue降风险Risk
梳理内容与使用情况,便于在诉讼未决背景下为授权定价。Catalogue content and usage so Torstar can price licences, given its pending lawsuits.建议 · 联合法务Proposed · with counsel
记者培训与实验空间Reporter training space
提效Efficiency体验Experience
为记者提供安全空间与短训,在真实文件上试用工具。A safe space and short training for reporters to test tools on real files.建议Proposed

12 个月路线12-month roadmap 建议Recommendation

奠基Foundation

第 1–3 月Months 1–3
  • 各类稿件工时基线Baseline hours per story type
  • 在用工具清单Inventory of tools in use
  • 点名流程负责人Process owner named
◆ 基线定稿Baseline final

试点Pilot

第 4–6 月Months 4–6
  • 一个版面使用工具One desk uses the tool
  • 对比测量Comparison measured
  • 首份节省工时报告First hours-saved report
◆ 指标达成Targets met

上线Go live

第 7–9 月Months 7–9
  • 工具向各版面开放Tool open to all desks
  • 编辑核验关键事实Editor checks key facts
  • 上线即监控Monitoring from day one
◆ 编辑部签字Editorial sign-off

扩展Extend

第 10–12 月Months 10–12
  • 流失预警试点Churn alert pilot
  • 授权立场初稿Licensing position drafted
  • 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

  • AI 辅助分析的事实须由编辑核验An editor checks facts from AI-assisted analysis
  • 记者所用工具全部建清单Inventory of every tool reporters use
  • 推荐类工具:写明目标与代理指标Recommendation tools: goal and proxy written down
  • 适用的版权与隐私规则,与法务确认Confirm applicable copyright and privacy rules with counsel

三层指标Three-layer metrics 建议Recommendation

采用Adoption每周使用共用工具的记者数Reporters using the shared tool weekly · 有活跃使用的版面数Desks with active use
过程Process每篇调查工时Hours per investigation · 编辑事实核验通过率Editor fact-check pass rate · 发稿前拦下的错误数Errors caught before publishing
结果Outcome对照基线节省的工时Hours saved against baseline · 订阅留存Subscriber retention

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

风险Risks 判断Judgment

只有作者会用的工具A tool only its author can run

记者自建的工具可能依赖某一个人;要写成文档并点名负责人。A reporter's tool may depend on one person; document it and name an owner.

错误见报Errors reach print

AI 辅助分析可能读错文件;关键事实须由编辑核验。AI-assisted analysis can misread a document; an editor must check key facts.

推荐里的代理指标Proxy targets in recommendations

模型可能优化点击而非读者信任;上线前写清两者。A model can optimise clicks, not reader trust; name both before launch.

前 90 天First 90 days 建议Recommendation

第 1–2 周Weeks 1–2点名流程负责人;定基线Name the process owner; fix baselines
第 3–4 周Weeks 3–4盘点记者当前所用工具Inventory tools reporters use today
第 5–10 周Weeks 5–10整理 Strong Mayors 流程;选定版面Document the Strong Mayors workflow; pick a desk
第 13 周Week 13首次复盘First review
今天只需要一个决定:One decision today: 指定全编辑室共用工具的负责人name the owner of the shared newsroom tool

我们可能错在:We could be wrong if: 其他版面没有类似的文件密集型工作,或工具节省的时间不足以支撑投入。other desks do not have similar document-heavy work, or the tool saves too little time to justify support.

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]