Overview › 档案Profiles › Individual Advisors

顶级个人 AI 顾问与演讲者 · 档案

对标"一把手 AI 转型闭门日"（一天一场 / 高密度小房间）的个人顾问与演讲者深度档案。 五人按"可借鉴度"排序：方法论可复制性、对一把手的说服力、产品形态契合度。 出场费均为演讲经纪机构公开估值区间，非本人公布价，仅作方向性参考。 调研截止：2026-06。

1. Ethan Mollick（★ 标杆 · 实境实战派）

一句话：Wharton 教授，《Co-Intelligence》作者，TIME 100 AI（2024）。靠"现场搭 AI agent"的实境式 keynote 让高管"亲眼看见"能力，是闭门日最值得照抄的范式。

背景履历 - Wharton 管理系副教授，Ralph J. Roberts 杰出学者 / Rowan Fellow；MIT Sloan 博士。 - Wharton Generative AI Labs 联合主任（与妻子 Lilach Mollick 共同主理，她负责教学法一侧）。 - TIME 100 AI 2024 入选。 - Substack「One Useful Thing」订阅者约 45 万+，是其影响力核心资产。

主打观点 / 著作 - 《Co-Intelligence: Living and Working with AI》（Portfolio / Penguin，2024.04，NYT 畅销书）。 - "锯齿前沿"（Jagged Frontier）：AI 能力不是平滑曲线，而是难以预测的锯齿边缘——看似同难度的任务有的在能力内、有的在能力外，只能靠亲手用才能摸清形状。 - Centaur vs. Cyborg：两种人机协作模式。Centaur = 按强项整块分工；Cyborg = 任务碎片来回交织。 - 《Co-Intelligence》四原则：① 永远把 AI 拉到桌上一起干（Always invite AI to the table）；② 做"环中之人"（human in the loop）；③ 把 AI 当人对待（但指定它是什么人/角色）；④ 假设这是你这辈子用过最差的 AI。 - "10 小时法则"：在真实工作里主动用前沿模型约 10 小时，直觉才会"咔哒"打开——这种直觉无法被代劳或被简报传授。 - （种子勘误：Mollick 没有"three sigma"提法；最接近的是他引用的 AI 家教"two sigma"效应。）

产品形态 - 招牌是实境演讲：在台上当场为客户真实业务搭一个能跑的 AI agent、几分钟生成一套幻灯片。客户证言："他展示了这些工具真正能干什么，当场搭了个 AI agent。" - 旗舰主题：From Disruption to Opportunity: Embracing the AI Revolution。 - 经纪：Washington Speakers Bureau（WSB）为主，多家 bureau 列名。个人站 moreintelligent.ai（当前 TLS 证书异常）。

出场费 - WSB 不公开报价（"直接联系"）。第三方 bureau 估值约 $75,000–$150,000/场，与种子区间一致——属指示性，非官方确认。

教学法 / 独特之处 - 实证研究背书是最大资产：BCG 田野实验（与 Harvard/MIT/Warwick + BCG，2023.09）——758 名顾问、18 项真实任务，用 GPT-4 者多完成 12.2% 任务、快 25.1%、质量高 40%；最弱的人提升最多（+43%，"拉平"效应）。这是全行业被引最多的 AI 生产力量化证据，也是他的信誉锚。 - 双重信誉：学术严谨 + 在任高管顾问（据 WSB 涉 JPMorgan、Google、白宫）。 - 方法 = 在前沿上边做边学，不是抽象战略；让高管"感受"能力，而非被描述。

对"一把手 AI 转型"的可借鉴点 - 用实境搭建开场，别用幻灯片：针对某真实业务问题当场搭一个能跑的 agent——对持怀疑态度的一把手是单点最强说服动作，直接照抄。 - 把"10 小时法则"定为领导班子的入门仪式；直觉不可外包。 - 用 BCG 数字（质量 +40% / 速度 +25% / 落后者 +43%）作为预算对话的实证 gate——这是该领域最经得起辩的数据点。 - Centaur/Cyborg 提供了"重设岗位"而非"加装工具"的词汇表。

局限 - 是学者/解释者，不是操盘手：他诊断与点燃，不替你跑转型、不搭你的技术栈、不背交付结果。 - 旗舰 keynote 高度可复制、同一套 demo 跑各场，不是定制深度顾问；你买的是"框架 + 哇感"，不是定制路线图。 - 通才，缺行业纵深与中国市场细节，例子偏美式知识工作。 - 内容会快速过时（他自己的"最差 AI"原则反噬其 demo）——框架长青，但具体技巧很快过期。

信息来源 / Sources

2. Cassie Kozyrkov（决策科学派 · 框架最干净）

一句话：前 Google 首席决策科学家（Chief Decision Scientist），训练过 2 万+ Googler。主张"AI 是决策学科，不是技术目的"——是给一把手房间纠偏 vendor hype 的最佳框架型声音。

背景履历 - Google 首位首席决策科学家：2014 入职（统计学家），2018 前后任此职，2023 离开（约 9–10 年）。在 Google 内部从零创立"决策智能（Decision Intelligence）"学科并运营约 5 年。 - 亲自培训 2 万+ Googler，助力 500+ 项目落地决策智能。 - 教育横跨统计、经济、心理、神经科学（神经经济学方向）；15 岁开始本科。 - 现为 Kozyr（2023 创立）创始人兼 CEO，联合 AI + 行为科学 + 战略。 - 自报触达：被 2 万+ CxO 关注；在 40+ 国家、七大洲演讲；HBS 高管 MBA 客座讲师；客户/组织含 Gucci、NASA、Spotify、Meta、Lenovo。

主打观点 / 框架 - 决策智能 = 把信息转化为行动的学科；AI/数据是决策的手段，不是技术终点。 - 决策科学 vs 数据科学：数据科学给答案，但"不覆盖人如何思考一个决策"；多数数据科学训练"假设决策者完全知道自己要什么、问题已被完美框定"——而那恰恰是真正的失败点。 - "如果你先看数据，你就做错了"（HBR, 2019.06）：先框定决策，再找数据；从"默认决策"（此刻没有更多证据你会选什么）出发。 - "第三类错误"：用全对的数学，解了完全错的问题。 - 2025 对 LLM/GenAI 立场：AI 产出无穷答案，但判断哪些有用、合伦理是深度人类任务；"AI 应是自动化的最后手段——只在传统方法搞不定的复杂问题上用"。 - AI 信任的"四支柱"：能力、正直、善意、魅力。

产品形态 - Keynote（最高曝光产品，世界最大舞台）、高管 workshop、在线视频课、1:1 顾问 / 顾问董事席位、Substack 通讯。 - "Making Friends with Machine Learning"（MFML）：原为 Google 内部课、训练 2 万+ Googler，现 YouTube 免费全集。（种子勘误：不是付费 Coursera 课。）

出场费 - AAE Speakers 报：线下 $50,000–$100,000，线上 $30,000–$50,000。与种子区间吻合。经纪机构阵容极广（AAE、BigSpeak、WSB、Harry Walker 等），是高需求信号。个人预订站 makecassietalk.com。

教学法 / 独特之处 - 决策优先、非技术优先——这是她区别于所有"AI 专家"的核心：先锚定决策与其框定，再谈模型/数据。 - 领导/C-suite 取向，概念重于数学（MFML 明确跳过数学与代码）。 - 翻译者角色：架接统计严谨 + 行为科学 + 高管语言。

对"一把手 AI 转型"的可借鉴点 - 定位契合度极高：她整套"AI 是决策学科 / 先框定决策再看数据 / AI 是自动化最后手段"正是一把手房间对 vendor hype 的纠偏。"默认决策 + 第三类错误"框架干净、好记、可直接复用。 - 信誉栈适配中国一把手的身份信号：Google 首位首席决策科学家 + 训练 2 万 Googler + HBS 讲师——对非技术 C-suite 房间最易读的 AI 领导力品牌。 - 形态匹配：本就做高管 workshop + 顾问，多小时闭门日在其标准产品线内。

局限 - 不是建造者 / 非在任操盘手：是教育者—布道者—顾问，深度在"决策框定与沟通"，不在前沿 ML、agent 或落地实现。想要"具体怎么搭/怎么部署"的房间，她是框架 keynote，不是技术架构师。 - 论点稳定但演进慢（决策智能讯息自 2018–19 大体一致）。 - 触达数据多为自报（2 万 CxO 关注、项目数等来自本人站/经纪页，非独立审计）；Wikipedia 本身挂有"独立来源不足"维护标记。 - 美国中心、仅英语，无中国履历，需翻译 + 文化本地化。

信息来源 / Sources

3. Zack Kass（前 OpenAI GTM · 内部人信誉 + 商业框架）

一句话：前 OpenAI Go-To-Market 负责人（2021–2023），AI 乐观主义未来学家。对一把手最有卖点的是"我在房间里待过"的内部人信誉 + 用商业/营收/颠覆语言讲 AI。

背景履历 - UC Berkeley 历史 + 计算机本科。早年 Figure Eight（前 CrowdFlower，ML 数据标注）销售/客户成功 VP；Lilt（AI 翻译）顾问/营收 VP。 - OpenAI Head of Go-To-Market，2021–2023（在 ChatGPT 发布前搭建销售/解决方案/合作团队）——这是各来源一致、最可核实的硬事实。 - 现任：UVA McIntire 商学院驻校高管；Ruder Finn AI 顾问委员会主席（2024 起）；Conservation International AI for Conservation Fellow；PRWeek "AI 25: Class of 2026"。 - （履历提示：其个人站对 Figure Eight / Lilt 给出重叠的 2017–2021 区间，"在 AI 14–16 年"为自报框架；干净锚点只有 OpenAI 2021–2023。）

主打观点 / 著作 - 著作《The Next RenAIssance: AI and the Expansion of Human Potential》（2026.01.13，USA Today 畅销书）。 - 品牌 = 技术乐观、丰裕（abundance）框架的 AI 未来主义，刻意对立于末日论。 - "无计量的智能（Unmetered Intelligence）"：AI 像电力/带宽一样变得丰裕近免费，人人默认获得博士级助手，"在智能丰裕的世界里，解决问题的成本坍塌"。 - "自动化边界（Automation Boundary）"（招牌概念）：社会"能自动化"与"选择保留为人类"之间的界线（照护、创造、精神指引）。"能自动化不代表应该自动化。" - 就业：历史决定论式乐观——强调"适应力"是 21 世纪核心技能（适应的时代）。 - 商业框架（GTM 视角）："如果你的价值主张建立在稀缺上……你会被颠覆。" - 他确实点名的风险偏温和：能力与预期错配、对开放式系统的信任构建、两端坏政策（过度监管 vs 监管不足）。

产品形态 - 主产品是 keynote；顾问为次、且产品化程度低。 - 个人站列：keynote（主）、advisory（表单式）、宽泛"战略咨询"。未列具体命名的 workshop / pilot 设计 / 治理 workshop 产品——这些非其打包产品。 - 招牌 keynote：The Next Renaissance、The Next Great Company、Societal Thresholds 等。

出场费 - 独家经纪 BigSpeak 多标"请询价"。二级 bureau 估值（仅作方向）：All American Speakers $30K–$50K；Key Speakers $30K–$65K；Gotham Artists $50K–$75K。工作估值 ~$30K–$75K/场，美国/线下取高端。闭门日顾问价不公开，需经 BigSpeak 直谈。

代表客户 - Fortune 500 偏多：Coca-Cola、Microsoft、Morgan Stanley、Samsung、TikTok、Deloitte、PayPal、MetLife、Broadcom、Amgen 等；自称亦服务"企业、非营利、政府"（政府客户未公开点名）。证言强调战略而非仅激励（可口可乐总裁："Zack 帮我们规划了整套 AI 战略，他就是懂。"）。

对"一把手 AI 转型"的可借鉴点 - 前 OpenAI GTM 信誉是对 C-suite/创始人最有营销力的单一资产——"内部人"。 - 说商业/战略/颠覆语言，不是安全研究黑话，落地营收导向的领导。 - "自动化边界"是真正可用的 workshop 框架：逼领导决定"自动化什么 vs 刻意保留人类什么"——对组织/产品战略直接可操作。 - 丰裕/乐观叙事给高管"有紧迫感但不瘫痪"的情绪弧线，适合开场/炉边。

局限 - 乐观即产品——风险处理轻，结构上是多头；不适合严肃的 AI 安全/治理/红队日。 - GTM/销售出身，非研究者或工程师；深度在商业化与叙事，不在模型内部、架构、落地。 - 产品是 keynote 主导、非产品化顾问；结构化闭门课程需定制，且非其核心肌肉。 - 溢价买一场"讲"，多数具体价值是激励 + 框架，ROI 取决于你自备执行能力配套。

信息来源 / Sources

4. Amy Webb（量化未来学派 · 情景规划方法论）

一句话：Future Today Strategy Group（前 Future Today Institute）创始人兼 CEO，NYU Stern 战略前瞻教授。卖的是可复制的前瞻方法论（信号→趋势→情景）与反炒作的严谨，而非预测本身。

背景履历 - FTSG 创始人兼 CEO（公司 2006 创立；2025.03.03 由 Future Today Institute 改名 Future Today Strategy Group，强调"前瞻与战略须协同"）。 - NYU Stern 战略前瞻兼任教授（自创 MBA 级课程）。 - 客户含 Fortune 100、三届白宫政府、美国联邦机构/国会、四大洲国家元首。 - 荣誉：Thinkers50 第 3 名最具影响力管理思想家；BBC 100 Women（2019）；CFR 终身会员。

主打观点 / 著作 - 四本书：《Data, A Love Story》（2013）、《The Signals Are Talking》（2016，方法论书）、《The Big Nine》（2019，美中九大 AI 公司 G-MAFIA + BAT）、《The Genesis Machine》（2022，合成生物学）。 - 年度科技趋势报告（旗舰资产）：自 2007 起，2025 版为第 18 版，约 1,000 页、其中约 107 页讲 AI；每年下载约 100 万次、免费，每年 SXSW 发布。 - 方法论：信号→趋势→情景。七步预测漏斗 + CIPHER（Contradictions / Inflections / Practices / Hacks / Extremes / Rarities）识别真信号；产出"可能/合理/可能性"（probable/plausible/possible）三层情景。核心："今日边缘是明日主流。" - 近期招牌：SXSW 2024「技术超级周期」（AI + 传感生态 + 生物科技三力收敛）；SXSW 2025「活的智能（Living Intelligence）」——能感知/学习/适应/演化的系统，语言模型→行动模型的转向。

产品形态 - Keynote（主公开产品）、FTSG 战略前瞻顾问/retainer（情景规划、颠覆识别、系统性风险/长期政策分析）、趋势 workshop（"像未来学家一样思考"）、免费年度报告作漏斗顶端引流。 - amywebb.io 是个人枢纽；商业引擎是 ftsg.com。

出场费 - bureau 报价分歧，宜报区间：VaynerSpeakers $40,001–$75,000（线下）；speakingfee.com $20,000–$50,000（2026 价），线上约 $12K–$17K。线下约 $20K–$75K，近期/溢价 listing 居 $40K+。种子的 $20K–50K 偏中低端。

教学法 / 独特之处 - 严谨、数据驱动、反炒作：被定位为"把战略前瞻变成严谨数据学科"，有显式方法（CIPHER + 漏斗 + 情景纪律），而非凭感觉。 - 信誉/记录：约 18 年年度报告、年下载百万；《The Big Nine》对 AI 集中度的框定大体应验；服务白宫、CFR 终身会员。

对"一把手 AI 转型"的可借鉴点 - 把 AI 放进更大的"超级周期 / 活的智能"收敛框架（AI + 传感 + 生物），而非孤立工具——契合一把手级战略视野。 - 给领导可复制的方法而非预测：漏斗 + CIPHER + 三层情景是可迁移的高管框架，"授人以渔"的闭门前瞻日格式。 - 把"像未来学家一样思考"练习 + 行业专属情景结合，逼领导班子直面 2030/2035 情景。 - 反炒作信誉是对怀疑型资深高管的卖点。

局限 - 宏观前瞻，非动手落地：产出是董事会/政策高度的战略/情景/趋势，不做 AI 工程、模型部署、数据基建或组织级执行；前瞻日定方向与心智模型，不产出 build plan / 技术栈 / 可交付能力。 - 收敛叙事宽，对单一技术是通才视角而非实践者深度。 - 溢价买 keynote/顾问形态，非 fractional-CTO/落地伙伴；最好与动手型 AI 落地顾问/团队配对。

信息来源 / Sources

5. William Flaiz（驻场密集型 · 最接近"闭门日"的形态原型）

一句话：前诺华数字战略高管，单干顾问。其"≤10 人、连续 5 天驻场、产 90 天路线图"密集营，是你"闭门日"概念最接近的形态/打包原型——但本人 AI 顾问业务几无公开 traction 证明。

背景履历（履历可核，AI 业务 traction 不可核） - 诺华：Executive Director, Web Experience & Strategy, International（2023.01–2024.05），主导 90 国 1,200 个遗留站点整合（"退役 900 站、运营成本降 52%"）——自报。 - Razorfish（约 2002–2011）：升至 GM, Global Solutions Group——MediaPost（2009）独立证实该晋升。 - 现为 CleanSmartLabs + Bottom Line Strategy Group 创始人（2024 起），做 AI 产品（CleanSmart 数据清洗等）。 - 教育：Drexel 信息系统硕士；MIT「Applied Generative AI for Digital Transformation」短期高管证书。 - 判读：他不是 AI/ML 技术领袖出身，而是数字营销 / martech / web 平台转型高管，2024 重新定位为"企业 AI"。

核心方法论 - 论点："73% 企业 AI 项目从未规模化超越试点"——框定为执行/战略问题而非技术问题。 - "四大基础"：① 上线前定义成功（业务指标非技术指标）；② 早期驾驭合规复杂度——招牌"合规优先 AI 战略"，把监管当设计约束而非障碍（源自 pharma/诺华经验）；③ 解决数据地基（数据准备占 60–80% 工作量，绑定其 CleanSmart 产品）；④ 把变革管理内建进战略。 - 四层架构：信号获取 → 数据完整性 → 智能引擎 → 决策系统。

产品形态 / 价格 - 招牌 = "Strategic AI Implementation Workshop"（5 天 AI 战略 workshop）：连续 5 天、线下、每场 ≤10 人，交付 90 天行动计划，后续 30 天咨询 + 季度回访；目标 12 个月内 15–30% 运营效率提升。 - 漏斗：免费"中市场 AI 战略评估"（AI 就绪分 + 90 天路线图）→ workshop。 - 无公开价，仅对标他价（企业咨询 $100K–$500K+ / DIY 18–24 月）。按同类单干 AI 顾问 ~$3K–$6K/天，5 天约 ~$15K–$50K（推断，非声明）。

对"一把手 AI 转型"的可借鉴点（他是最接近的分析原型） - 格式：密集、限员（≤10）、驻场、强同侪互动——验证"小房间高密度"前提。但他是 5 天驻场营，比你单日闭门重得多；单日版是更轻的不同产品。 - 交付物：90 天路线图作为有形产物是关键销售钩——领导买的是具体计划而非激励，直接照抄。 - 定位："Fortune-500 打法 + 中市场现实"、合规优先、"试点→生产"——强且可复制的框架。 - 漏斗设计：免费 AI 就绪评估 → 路线图 → workshop，两步漏斗值得抄；30 天 + 季度回访把一次性转为 retainer。

局限 + 可核实性警示（关键） - 可独立核实：身份、Razorfish GM 晋升（MediaPost 2009）、Drexel/MIT 教育、诺华 ED 角色类型。 - 仅自报、需谨慎：所有亮眼数字（52% 降本、$24M→$50M、11X 回报等）只出现在其营销页且页间不一致（如某处称诺华省 ">$200M"，自家页强调"52%"）——营销夸大黄旗。 - 薄/不可核：workshop 定价无公开、AI workshop 客户无点名、该驻场是否真跑过无证言/案例。当前 AI 顾问业务全部回溯到其自家站/Medium，无任何第三方评论、客户证言或独立案例。 - 结构判读：这是产品化服务 / 内容漏斗打法（Medium 高产 + 评估引流 + 高客单 workshop），由可信但单干的操盘手运营。是好的形态模板/打包模板，不是竞争威胁，也非高天花板生意。其历史客户（Gates Foundation/ICANN/Microsoft）是 2002–2017 的 agency 项目，非 AI 顾问客户，挪用为"客户"会失实。

信息来源 / Sources

---

Top Individual AI Advisors & Speakers

Deep profiles of individual advisors/speakers benchmarked against a "CEO AI-transformation closed-door day" (one engagement/day, high-density small room). Five people ranked by "borrowability": replicability of method, persuasiveness to top decision-makers (一把手), and fit of product form. All speaking fees are public speaker-bureau estimate ranges, not self-published rates — directional only. Research cutoff: June 2026.

1. Ethan Mollick (★ benchmark · live experiential)

One line: Wharton professor, author of Co-Intelligence, TIME 100 AI (2024). His "build an AI agent live on stage" experiential keynote makes executives see the capability — the single most copyable template for a closed-door day.

Background - Associate Professor of Management, Wharton; Ralph J. Roberts Distinguished Faculty Scholar / Rowan Fellow; PhD, MIT Sloan. - Co-director of Wharton's Generative AI Labs (with wife Lilach Mollick, who leads the pedagogy side). - TIME 100 AI 2024. - "One Useful Thing" Substack: ~450k+ subscribers — his core influence asset.

Core ideas / books - Co-Intelligence: Living and Working with AI (Portfolio/Penguin, Apr 2024, NYT bestseller). - The Jagged Frontier: AI capability is not a smooth curve but an unpredictable jagged edge — equally hard-looking tasks fall inside or outside competence; you only learn its shape by using it. - Centaur vs. Cyborg: two human-AI modes. Centaur = clean division of labor by strength; Cyborg = fragments handed back and forth mid-task. - The four principles: ① Always invite AI to the table; ② Be the human in the loop; ③ Treat AI like a person (but tell it which kind); ④ Assume this is the worst AI you'll ever use. - The 10-hour rule: ~10 hours of active use on real work is when intuition clicks — it can't be delegated or briefed. - (Seed correction: there is no "three sigma" of his; the closest is the "two sigma" AI-tutor effect he cites.)

Product form - Signature = experiential keynotes: building a working AI agent for the client's real business live, generating a deck in minutes. Client testimonial: "he showed us what these tools can actually do, building an AI agent live." - Flagship: From Disruption to Opportunity: Embracing the AI Revolution. - Agency: Washington Speakers Bureau (WSB) primary; listed across bureaus. Personal site moreintelligent.ai (currently has a TLS cert issue).

Speaking fee - WSB does not publish ("contact directly"). Third-party bureau estimates: $75,000–$150,000/engagement, consistent with the seed — indicative, not officially confirmed.

Pedagogy / what's unique - Empirical research base is the headline asset: the BCG field experiment (with Harvard/MIT/Warwick + BCG, Sept 2023) — 758 consultants, 18 real tasks; GPT-4 users completed 12.2% more tasks, 25.1% faster, 40% higher quality; lowest performers gained most (+43%, the "leveling" effect). The most-cited quantitative proof of AI productivity, and his credibility anchor. - Dual credibility: academic rigor + active executive advising (per WSB: JPMorgan, Google, White House). - Method = learning by doing on the frontier; he makes executives feel capability rather than describe it.

Takeaways for a CEO AI-transformation day - Lead with a live build, not slides — building a working agent on a real problem is the single strongest move for skeptical decision-makers. Copy this directly. - Make the 10-hour rule the leadership team's entry ritual; intuition can't be outsourced. - Use the BCG numbers (+40% quality / +25% speed / +43% for laggards) as the empirical gate for budget conversations — the most defensible data point in the field. - Centaur/Cyborg gives vocabulary for redesigning roles, not just bolting on tools.

Limitations - An academic/explainer, not an operator — diagnoses and inspires; doesn't run your transformation, build your stack, or own outcomes. - Highly repeatable flagship keynote with the same demos across audiences — not bespoke deep advisory; you buy the framing + the "wow," not a custom roadmap. - Generalist; light on industry depth and China-market nuance; examples skew US knowledge work. - Content dates fast (his own "worst AI" principle bites his demos) — frameworks endure, specific tricks expire.

信息来源 / Sources

2. Cassie Kozyrkov (decision-science school · cleanest frameworks)

One line: Former Google Chief Decision Scientist, trained 20,000+ Googlers. Thesis: "AI is a decision discipline, not a tech end" — the best framing voice to correct vendor hype in a decision-maker room.

Background - Google's first Chief Decision Scientist: joined 2014 (as a statistician), held the role ~2018, left 2023 (~9–10 yrs). Founded the "Decision Intelligence" discipline inside Google from scratch and ran it ~5 years. - Personally trained 20,000+ Googlers; helped 500+ projects adopt decision intelligence. - Education spans statistics, economics, psychology, neuroscience (neuroeconomics); started undergrad at 15. - Now Founder & CEO of Kozyr (2023), uniting AI + behavioral science + strategy. - Self-reported reach: followed by 20,000+ CxOs; spoken in 40+ countries, all 7 continents; HBS executive-MBA guest lecturer; clients/orgs include Gucci, NASA, Spotify, Meta, Lenovo.

Core ideas / frameworks - Decision Intelligence = the discipline of turning information into action; AI/data is the means, not the tech end. - Decision science vs data science: data science gives answers but "doesn't cover how humans think through a decision"; most training "assumes the decision-maker knows exactly what they need and the problem is framed perfectly" — the actual failure point. - "If you look at data first, you're doing it wrong" (HBR, 2019.06): frame the decision before seeking data; start from a "default decision." - "Type Three Error": all the right math to solve entirely the wrong problem. - 2025 stance on LLMs/GenAI: AI produces endless answers, but judging which are useful and ethical is a deeply human task; "AI should be a last resort for automation." - "Four Pillars of Trust" for AI: competence, integrity, benevolence, charisma.

Product form - Keynotes (highest-profile product), executive workshops, online video courses, 1:1 advisory / advisory-board seats, Substack. - "Making Friends with Machine Learning" (MFML): originally an internal Google course (trained 20k+ Googlers), now free in full on YouTube. (Seed correction: not a paid Coursera course.)

Speaking fee - AAE Speakers: in-person $50,000–$100,000, virtual $30,000–$50,000 — matches the seed. Broad bureau roster (AAE, BigSpeak, WSB, Harry Walker) signals high demand. Personal booking site: makecassietalk.com.

Pedagogy / what's unique - Decision-first, not tech-first — the core differentiator vs every "AI expert": anchor on the decision and its framing before any model/data. - Leadership/C-suite orientation, conceptual over mathematical (MFML explicitly skips math and code). - Translator role: bridges statistical rigor + behavioral science + executive language.

Takeaways for a CEO AI-transformation day - Strongest positioning fit: her thesis is exactly the corrective a decision-maker room needs vs hype. "Default decision" + "Type Three Error" are clean, memorable, copy-pastable. - Credential stack fits status-signaling: Google's first Chief Decision Scientist + 20k Googlers trained + HBS lecturer — the most legible AI-leadership brand for a non-technical C-suite room. - Format fit: does executive workshops + advisory, so a multi-hour closed-door day is within her product line.

Limitations - Not a builder / not a current operator — educator-evangelist-advisor; depth is decision framing & communication, not frontier ML, agents, or implementation. For "how do we actually build/deploy," she's the framing keynote, not the architect. - Stable thesis, slow evolution (consistent since 2018–19). - Reach metrics largely self-reported; Wikipedia itself carries "insufficient independent sourcing" maintenance flags. - US-centric, English-only, no China track record — needs interpretation + localization.

信息来源 / Sources

3. Zack Kass (ex-OpenAI GTM · insider credibility + business framing)

One line: Former OpenAI Head of Go-To-Market (2021–2023), AI-optimist futurist. His top selling point to decision-makers: "I was in the room" insider credibility + speaking AI in business/revenue/disruption language.

Background - UC Berkeley History + CS. Early career: Figure Eight (ex-CrowdFlower, ML data-labeling) sales/CS VP; Lilt (AI translation) advisor/revenue VP. - OpenAI Head of Go-To-Market, 2021–2023 (built sales/solutions/partnerships teams before ChatGPT launch) — the consistent, verifiable hard fact. - Now: Executive-in-Residence, UVA McIntire; Chair, Ruder Finn AI Advisory Council (2024–); Conservation International AI-for-Conservation Fellow; PRWeek "AI 25: Class of 2026." - (CV note: his site gives overlapping 2017–2021 ranges for Figure Eight/Lilt; "14–16 years in AI" is self-framing; the clean anchor is only OpenAI 2021–2023.)

Core ideas / book - Book: The Next RenAIssance: AI and the Expansion of Human Potential (Jan 13, 2026, USA Today bestseller). - Brand = techno-optimist, abundance-framed AI futurism, deliberately anti-doomer. - "Unmetered Intelligence": AI becomes abundant and near-free like electricity/bandwidth; everyone gets PhD-level assistance by default; "the cost of problem solving collapses." - "Automation Boundary" (signature): the line between what societies can automate and what they choose to keep human (caregiving, creativity, spiritual guidance). "Just because we can automate something doesn't mean we should." - Jobs: historical-determinist optimism — "adaptability" as the core 21st-century skill (the Age of Adaptation). - Business framing (GTM lens): "If your value proposition is based on scarcity… you're going to be disrupted." - Named risks are notably milder: capability-vs-expectation misalignment, trust with open-ended systems, bad policy on both ends.

Product form - Primary product is keynote; advisory is secondary and lightly productized. - His site lists keynotes (primary), advisory (contact form), broad "strategic consulting." No discrete named workshop / pilot-design / governance products. - Signature keynotes: The Next Renaissance, The Next Great Company, Societal Thresholds, etc.

Speaking fee - Exclusive bureau BigSpeak mostly "please inquire." Secondary-bureau estimates (directional): All American Speakers $30K–$50K; Key Speakers $30K–$65K; Gotham Artists $50K–$75K. Working estimate ~$30K–$75K/keynote, US/in-person at the high end. Closed-door advisory rate not public — negotiate via BigSpeak.

Representative clients - Fortune 500-heavy: Coca-Cola, Microsoft, Morgan Stanley, Samsung, TikTok, Deloitte, PayPal, MetLife, Broadcom, Amgen, etc.; claims businesses/nonprofits/governments (government clients not named). Testimonials stress strategy, not just inspiration (Coca-Cola President: "Zack helped us plan our entire AI strategy. He just gets it.").

Takeaways for a CEO AI-transformation day - Ex-OpenAI GTM credential is the most marketable single asset for a C-suite/founder room — "someone who was inside." - Speaks business/strategy/disruption, not safety-research jargon — lands with revenue-minded leaders. - "Automation Boundary" is a genuinely usable workshop framework: forces leaders to decide what to automate vs deliberately keep human — directly actionable. - Abundance/optimism gives executives "urgency without paralysis" — good opener/fireside arc.

Limitations - Optimism is the product — light on risk, structurally a bull; not for a sober AI-safety/governance/red-team day. - GTM/sales background, not researcher/engineer; depth in commercialization and narrative, not model internals or implementation. He frames; he doesn't build. - Keynote-led, not productized advisory; a structured closed-door curriculum must be custom-scoped and isn't his core muscle. - Premium price for a talk; most concrete value is inspiration + framing; ROI depends on execution capacity you supply.

信息来源 / Sources

4. Amy Webb (quantitative-futurist school · scenario-planning methodology)

One line: Founder & CEO of Future Today Strategy Group (formerly Future Today Institute), strategic-foresight professor at NYU Stern. Sells a replicable foresight methodology (signals→trends→scenarios) and anti-hype rigor, not predictions.

Background - Founder & CEO of FTSG (firm founded 2006; renamed Future Today Institute → Future Today Strategy Group on 2025.03.03, signaling "foresight and strategy must work in harmony"). - Adjunct Professor of Strategic Foresight, NYU Stern (built her own MBA-level course). - Clients: Fortune 100, three White House administrations, US federal agencies/Congress, heads of state on four continents. - Honors: Thinkers50 #3 most influential management thinker; BBC 100 Women (2019); CFR life member.

Core ideas / books - Four books: Data, A Love Story (2013), The Signals Are Talking (2016, the methodology book), The Big Nine (2019, the nine US/China AI giants — G-MAFIA + BAT), The Genesis Machine (2022, synthetic biology). - Annual Tech Trends Report (flagship): since 2007; the 2025 edition is the 18th, ~1,000 pages, ~107 on AI; downloaded ~1 million times/year, free, launched at SXSW. - Method: signals→trends→scenarios. Seven-step Forecasting Funnel + CIPHER (Contradictions / Inflections / Practices / Hacks / Extremes / Rarities) to spot true signals; outputs probable/plausible/possible scenarios. Core: "today's fringe is tomorrow's mainstream." - Recent headline theses: SXSW 2024 "Technology Supercycle" (convergence of AI + sensor ecosystems + biotech); SXSW 2025 "Living Intelligence" — systems that sense/learn/adapt/evolve; shift from language models → action models.

Product form - Keynotes (primary public product), FTSG strategic-foresight advisory/retainer (scenario planning, disruption identification, systemic-risk/long-term policy analysis), trend workshops ("think like a futurist"), the free annual report as top-of-funnel lead magnet. - amywebb.io is the personal hub; the commercial engine is ftsg.com.

Speaking fee - Bureaus disagree — quote a band: VaynerSpeakers $40,001–$75,000 (in-person); speakingfee.com $20,000–$50,000 (2026 rates), virtual ~$12K–$17K. In-person ~$20K–$75K, recent/premium listings center $40K+. The seed's $20K–50K is the low-to-mid end.

Pedagogy / what's unique - Rigorous, data-driven, anti-hype — positioned as having turned strategic foresight into a rigorous data discipline with an explicit method (CIPHER + funnel + scenario discipline), not vibes. - Track record: ~18 years of annual reports, ~1M downloads/yr; The Big Nine's AI-concentration framing largely held; advises the White House; CFR life member.

Takeaways for a CEO AI-transformation day - Frame AI inside the broader "supercycle / Living Intelligence" convergence (AI + sensors + biotech), not as a standalone tool — fits a decision-maker's strategic horizon. - Give leaders a repeatable method, not predictions: funnel + CIPHER + three-tier scenarios is a transferable executive framework — a "teach them to fish" foresight-day format. - Pair "think like a futurist" exercises with industry-specific 2030/2035 scenarios to force the leadership team to confront the future. - Anti-hype credibility is the selling point to skeptical senior executives.

Limitations - Macro foresight, not hands-on implementation: output is board/policy-altitude strategy/scenarios/trends; she doesn't do AI engineering, model deployment, data infra, or org-level execution. A foresight day sets direction and mental models, not a build plan, stack, or shipped capability. - Convergence narratives are wide — a generalist's view of AI, not a practitioner's depth. - Premium price for keynote/advisory; best paired with a hands-on AI implementation advisor/team to convert scenarios into a roadmap.

信息来源 / Sources

5. William Flaiz (intensive residency · closest form prototype to a "closed-door day")

One line: Former Novartis digital-strategy executive, solo advisor. His "≤10 people, 5 consecutive days on-site, produces a 90-day roadmap" residency is the closest form/packaging prototype to your "closed-door day" concept — but his AI-advisory business has almost no public traction proof.

Background (CV verifiable; AI-business traction not) - Novartis: Executive Director, Web Experience & Strategy, International (Jan 2023–May 2024), led consolidation of 1,200 legacy sites across 90 countries ("retired 900 sites, cut operating costs 52%") — self-reported. - Razorfish (~2002–2011): rose to GM, Global Solutions Group — independently confirmed by MediaPost (2009). - Now founder of CleanSmartLabs + Bottom Line Strategy Group (2024–), shipping AI products (CleanSmart data-cleaning, etc.). - Education: M.S. Information Systems, Drexel; MIT "Applied Generative AI for Digital Transformation" short exec certificate. - Read: not an AI/ML technical leader by training — a digital-marketing / martech / web-platform transformation executive who repositioned into "enterprise AI" in 2024.

Core methodology - Thesis: "73% of enterprise AI initiatives never scale beyond pilot" — framed as an execution/strategy problem, not a tech problem. - "Four Critical Foundations": ① define success before launch (business not technical metrics); ② navigate regulatory complexity early — his "Compliance-First AI Strategy," treat regulation as a design constraint not a barrier (from the pharma/Novartis experience); ③ solve the data-foundation problem (data prep = 60–80% of effort, tied to CleanSmart); ④ build change management into strategy. - Four-layer architecture: Signal Acquisition → Data Integrity → Intelligence Engines → Decision Systems.

Product form / pricing - Signature = "Strategic AI Implementation Workshop" (5-Day AI Strategy Workshop): 5 consecutive in-person days, ≤10 participants, deliverable a 90-day action plan, follow-up 30-day consultation + quarterly check-ins; target 15–30% operational-efficiency gain within 12 months. - Funnel: free "Mid-Market AI Strategy Assessment" (AI-readiness score + 90-day roadmap) → workshop. - No public price — only framed against alternatives (enterprise consulting $100K–$500K+; DIY 18–24 months). At ~$3K–$6K/day for comparable solo AI consultants, 5 days implies ~$15K–$50K (inferred, not stated).

Takeaways for a CEO AI-transformation day (the closest analog) - Format: intensive, capped cohort (≤10), on-site, strong peer interaction — validates the "small-room, high-density" premise. But his is a 5-day residency, far heavier than a single day; a one-day version is a lighter, different product. - Deliverable: the 90-day roadmap as the tangible artifact is the key sales hook — leadership pays for a concrete plan, not inspiration. Copy this. - Positioning: "Fortune-500 playbook for mid-market realities," compliance-first, "pilots → production" — strong, copyable framing. - Funnel design: free AI-readiness assessment → roadmap → workshop is a clean two-step funnel worth copying; 30-day + quarterly check-ins turn a one-off into a retainer.

Limitations + verifiability warning (critical) - Independently verifiable: identity, the Razorfish GM promotion (MediaPost 2009), Drexel/MIT education, the Novartis ED role type. - Self-reported only (caution): all the punchy numbers (52% cost cut, $24M→$50M, 11X return) appear only on his marketing surfaces and vary between pages (one cites Novartis savings ">$200M" while his own pages emphasize "52%") — a marketing-inflation yellow flag. - Thin / unverifiable: workshop pricing not public, no named AI-workshop clients, no testimonials/case studies proving the residency has run. The current AI-advisory business traces entirely to his own site/Medium; no third-party review found. - Structural read: a productized-services / content-funnel play (heavy Medium output + assessment lead magnet + high-ticket workshop) run by a credible-but-solo operator. A good format/packaging template, not a competitive threat or high-ceiling business. His historical clients (Gates Foundation/ICANN/Microsoft) are 2002–2017 agency engagements, not AI-advisory clients — borrowing them as "clients" would misstate.

信息来源 / Sources