Montai CEO Georgiadis Warns AI Startups Risk Losing 40% of Top Talent

Fireside Chat: Women in AI Leadership With Fortune And Therapeutics — Ai4

Ai4

AI firms that rely strictly on massive pay packages rather than inclusive, high-autonomy cultures will forfeit up to 40% of premier technical talent.

AI-assisted summary of Fireside Chat: Women in AI Leadership With Fortune And Therapeutics — original source. Check the original for full context.

Source published: 2025-08-25

Detailed notes

## Background Published on YouTube by the Ai4 channel on August 25, 2025, this video documents a fireside chat between Emma Hinchliffe, Editor at Fortune Most Powerful Women, and Margo Georgiadis, Co-Founder & CEO of Montai Therapeutics. According to the page description, the session provides a candid exploration of women's leadership journeys, insights, and impact on the future of artificial intelligence.

## TL;DR Margo Georgiadis discusses Montai Therapeutics' mission to use AI to address chronic disease and explains why defensibility in the AI era depends on distribution and engineering culture rather than compensation alone. She details the structural barriers and pervasive bro culture facing female technical talent, underscoring how algorithmic bias and lack of representation directly jeopardize product efficacy. Georgiadis offers practical strategies for inclusive hiring, closing the female self-advocacy gap, and urging executives to commit to hands-on AI upskilling to lead organizational change.

## Atmosphere & style The discussion has an intimate, candid, and empowering atmosphere, marked by a warm yet direct rapport between Hinchliffe and Georgiadis in front of an engaged audience. Georgiadis communicates with pragmatic authority and relatable humor, grounding complex tech strategy in vivid personal anecdotes about interview red flags and board-level realities. The session feels purposeful and energizing, shifting from sharp critiques of tech industry pitfalls into an inspiring, actionable call to champion inclusive leadership.

## Key takeaways - [00:04](https://www.youtube.com/watch?v=wDLGYWBalvs&t=4s) Montai Therapeutics leverages AI to develop small molecule therapeutics for chronic diseases, targeting an access gap where fewer than 10% of diagnosed patients receive transformative treatments. - [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) As foundation models encroach on specialized vertical software, defensibility and resilience for AI startups depend on distribution and becoming an entrenched application of record. - [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) Winning the war for top AI engineering talent requires a culture of trust, autonomy, and shipping production code immediately rather than relying solely on outsized compensation packages. - [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) Pervasive 'bro culture' and performative tests of dedication—such as demanding 24/7 on-call availability—alienate roughly 70% of women in tech who prioritize authentic impact. - [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) Tech companies should dismantle hiring bias by rewriting job descriptions around essential complementary skills and using data-driven audits to identify biased interview scoring patterns. - [06:38](https://www.youtube.com/watch?v=wDLGYWBalvs&t=398s) Diverse representation among AI architects and leaders is essential to prevent embedding historical biases into core systems that restructure work, society, and human outcomes. - [09:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=565s) Significant historical data gaps—such as only 1% of research historically targeting women's diseases—highlight why inclusion must be treated as a core business OKR to ensure models function safely for all users. - [09:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=565s) B2B AI adoption in highly specialized fields like life sciences faces deep cultural resistance from domain experts who fear their hard-won specialized knowledge will be made obsolete. - [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) Dismissing diversity during fast-paced market cycles alienates 30% to 40% of the elite talent pool, while a lack of visible female leadership causes rising women to systematically opt out. - [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) Because men disproportionately advocate for their own promotions, leaders must consciously spot unvoiced talent, while women must proactively put themselves in the path of opportunity. - [16:22](https://www.youtube.com/watch?v=wDLGYWBalvs&t=982s) A pronounced AI confidence and training divide persists: roughly 80% of professional women report inadequate AI education, and only 30% feel comfortable with AI compared to 71% of men. - [16:22](https://www.youtube.com/watch?v=wDLGYWBalvs&t=982s) To effectively guide enterprise AI adoption, C-suite executives need to dedicate two to three hours every week to direct, hands-on immersion and stress-testing of models. - [18:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=1105s) Overcoming deep-seated organizational resistance to AI requires resetting success metrics and empowering empathetic 'change integrators' to guide staff through operational transitions.

## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=wDLGYWBalvs&t=4s) Introduction and Montai Therapeutics' Mission

Emma Hinchliffe opens the session by welcoming Margo Georgiadis to the stage, noting the packed room of attendees and women leaders at the Ai4 conference. Highlighting Georgiadis’s diverse background—having previously served as the CEO of Ancestry, held leadership roles at Google, and sat on the boards of major corporations like McDonald’s and Mattel—Hinchliffe asks her to introduce her current venture, Montai Therapeutics.

Georgiadis explains that Montai is driven by an obsession with developing small molecule therapeutics to tackle chronic disease. Despite rapid innovation across the life sciences sector, transformative treatments fail to reach the vast majority of patients. Georgiadis points out a stark disparity: between 30% to 50% of the most severely diagnosed patients, and fewer than 10% of all diagnosed patients overall, lack access to life-changing medicines. To bridge this divide, Montai uses AI to reimagine drug discovery from first principles. Though transitioning into biotech was a steep learning curve given her non-life sciences background, she highlights that the company is already preparing to advance its first four therapies into the clinic.

- **Montai Therapeutics Focus:** Small molecule therapeutics targeting chronic diseases. - **Healthcare Access Gap:** Less than 10% of all diagnosed patients, and 30%–50% of severely diagnosed patients, currently have access to transformative therapies. - **Clinical Pipeline:** Advancing its first 4 drug candidates toward clinical trials.

---

[01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) AI Startups: The Application Layer, Distribution, and Talent Culture

Turning to the broader technological landscape, Hinchliffe asks Georgiadis to contrast the challenges facing legacy corporations integrating AI against those encountered by AI-native startups. Georgiadis describes the current landscape among foundation model builders and specialized software applications as "just a war." As foundation models threaten to encroach on vertical tools, founders are increasingly moving up the stack to the application layer. In her view, long-term resilience and defensibility will come down to distribution. Gaining a foothold as the entrenched "application of record"—particularly in high-stakes fields like healthcare where users are acutely sensitive to small differences—creates powerful moats that are difficult to displace.

```mermaid flowchart TD accTitle: AI Competitive Landscape and Resiliency Factors subgraph MarketDynamics["Market Pressures"] FM["Foundation Model Layer\n(Commoditization & Aggressive Expansion)"] AL["Application Layer\n(Vertical Specialized Tools)"] FM -.->|"Encroaches upon"| AL end

subgraph DefensibilityMoats["Defensibility & Resilience"] DIST["Distribution Foothold\n('Application of Record')"] CULT["Culture & Meaningful Impact\n(Retaining Top Engineering Talent)"] end

AL --> DIST AL --> CULT ```

Beyond competitive moats and distribution, Georgiadis identifies the fierce war for engineering talent as the defining challenge for AI startups. While the industry is dominated by massive compensation packages, she argues that leading-edge companies are realizing compensation alone is insufficient. Drawing on her board and advisory roles, she cites AppLovin—which she characterizes as "probably the biggest $150 billion company you've never heard of"—as a prime example of building an engineering-first culture. Rather than simply outbidding rivals, AppLovin empowers young engineers and interns to build and ship production code immediately. This culture of trust and tangible impact has allowed them to lure top-tier talent away from competitors like Grok and Tesla, demonstrating that high performers, particularly women, prioritize belonging and meaningful contribution over pure pay packages.

- **Defensibility Key:** Distribution and becoming the indispensable application of record in sensitive domains like healthcare. - **Talent Retention Strategy:** Moving beyond outsized pay packages to provide autonomy, fast shipping cycles, and authentic recognition. - **Intern Impact:** Interns working on production features and shipping real-world code rather than administrative tasks.

---

[04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) Reimagining Tech Culture and Inclusive Hiring to Attract Women

Hinchliffe brings up recent developments around Mira Murati and Thinking Machines Lab, where team members turned down massive multi-million and billion-dollar packages, asking what this reveals about how women build in AI and what they prioritize. Georgiadis confirms that culture plays an enormous role, sharing that in her mentorship of rising female technical leaders, she routinely hears shocking interview horror stories.

She recounts a recent incident where a premier female product and engineering leader interviewed with a leading foundation model company. The hiring manager bluntly told her, "And you're going to have to be on a beeper 24 hours a day, seven days a week," assuming this aggressive framing would test her dedication. Off-putting and tone-deaf to what actually drives high performers, the remark caused the candidate to end the interview on the spot. Georgiadis emphasizes that this heavy-handed posturing reflects the pervasive "bro culture sexism" experienced by roughly 70% of women in tech, who feel excluded if they do not fit the mold of pulling extreme hours or "slinging beers at the bar afterwards."

```mermaid flowchart LR accTitle: Strategies for Inclusive AI Talent Pipelines A["Job Descriptions"] -->|"Rewrite for broad skills & inclusivity"| B["Candidate Pipeline"] B -->|"Deploy AI auditing"| C["Interviewer Slates"] C -->|"Filter out biased interviewers"| D["Diverse & High-Impact Hires"] ```

To systematically fix this dynamic, Georgiadis urges tech leaders to redesign their hiring pipelines: - **Inclusive Job Descriptions:** Rewriting requisitions to focus on essential, complementary skills rather than homogenous profiles. - **Data-Driven Interview Audits:** Analyzing interview scoring trends to uncover bias, recalling her tenure at Google where data revealed specific interviewers consistently failed diverse candidates and candidates of color regardless of merit. - **Deploying AI for Hiring Equity:** Utilizing AI to evaluate pipelines and counter pattern-matching biases.

---

[06:38](https://www.youtube.com/watch?v=wDLGYWBalvs&t=398s) Cultivating Human-Centered Impact and Diverse Leadership in AI Architecture

Addressing how to pitch mission and purpose in sectors without an obvious humanitarian focus like biotech, Georgiadis reiterates that impact is about how organizations treat their people. In high-scale tech firms like AppLovin, which functions essentially as a massive AI and ad-serving engine, impact translates into working in a flat, non-hierarchical environment where individual creativity is celebrated and executive leadership genuinely cares for employee well-being. She shares an anecdote of an engineering leader who instituted a mandatory four-day holiday for an international team specifically requiring employees to bring their parents, respecting the familial values of that local culture.

Hinchliffe asks why female representation in AI remains an urgent priority rather than a secondary concern in a fast-paced market war. Georgiadis asserts that the industry is at a historic inflection point: AI is actively restructuring work, economies, and everyday life. Because how products are designed and architected fundamentally alters human outcomes, diversity in the room is critical to prevent repeating historical mistakes. She highlights the leadership of figures like Mira Murati championing open source, as well as Fidji Simo, whose work bridges technical capabilities with broader human outcomes. Drawing on decades of collective experience in the room, Georgiadis reminds the audience of past algorithmic failures—ranging from biased predictive policing models to flawed healthcare algorithms—warning that technology cannot achieve safe and widespread adoption if it fails to represent the full breadth of the population it serves.

[09:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=565s) Algorithmic Bias, Inclusion as a Business Imperative, and B2B vs. B2C Challenges

Addressing historical disparities in data, Margo Georgiadis points out that only 1% of research has historically focused on women-specific diseases. Similar data biases appear across clinical trials for underrepresented groups lacking regular healthcare access, as well as in algorithmic recruiting tools where automated bots simply clone existing employee profiles from past résumés rather than evaluating a broader spectrum of capabilities. Leading organizations do not ignore these risks; instead, they establish concrete OKRs and treat diversity as an essential business imperative to ensure their AI products perform reliably for everyone.

Drawing on her previous tenure as CEO of Ancestry, Georgiadis illustrates how algorithmic development requires diverse perspectives at the design table. When Ancestry developed AI-powered family history tools and health offerings, product teams had to account for how people from different ethnic backgrounds and immigrant histories would uniquely experience the product, user interface, and marketing. She stresses that inclusivity is just as critical in B2B environments as in consumer-facing applications. In life sciences, for instance, specialists who have spent twenty years mastering a single cell type often exhibit strong cultural resistance to multifactorial, integrated AI tools out of fear that their expertise will be made obsolete.

```mermaid flowchart TD accTitle: Systemic AI Bias and Integration Challenges

subgraph DataRisks["Data & Algorithmic Biases"] A["Biased Training Sets (e.g., 1% Women's Health Research)"] --> C["Flawed AI Models & Biased Recruiting Bots"] B["Homogeneous Legacy Profiles"] --> C end

subgraph SolutionStrategy["Enterprise Imperatives"] D["Diverse Teams & Product OKRs"] E["Empathetic UI & Inclusive Experience Design"] end

subgraph AdoptionBarriers["B2B & B2C Adoption"] F["Specialized Workforce Resistance (e.g., Life Sciences Domain Experts)"] G["Consumer Trust & Cultural Sensitivity"] end

C --> D D --> E E --> F E --> G ```

- Historically, only 1% of scientific research has targeted diseases specific to women. - Algorithmic hiring systems trained on historical résumés replicate existing workforce demographics rather than evaluating holistic skills. - High specialization in B2B fields like life sciences breeds resistance to AI adoption unless organizations proactively manage workflow integration and retraining.

---

[11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) Making the Business Case to Skeptics and Closing the Self-Advocacy Gap

When asked by Emma Hinchliffe how to counter executives who view the current AI race as an all-hands-on-deck "war moment" with no time to prioritize diversity, Georgiadis emphasizes the direct talent calculation: ignoring inclusion means intentionally walking away from 30% to 40% of the most talented, dedicated contributors. She has candid conversations with male-dominated leadership teams, warning them that if women see no female representation at the executive level, they systematically opt out of the company because they see no viable career trajectory.

Georgiadis also highlights a pronounced behavioral disparity between male and female employees regarding career advancement. Male employees regularly walk into executive offices to assert their readiness for promotions or pitch themselves for exciting new assignments, whereas women do so far less frequently. She urges leaders to be conscious of this cultural difference so they do not default to promoting only those who aggressively lobby for opportunities. At the same time, she challenges women in the room to proactively place themselves in the path of opportunity:

> "Remind yourself that if you don't put yourself in the path of opportunity and let people know what you're interested in and ask them, 'Do you think I'm qualified? And if I'm not, what do you think I need to do?' If you don't make that effort, you don't remind them that you want that."

- Disregarding diversity alienates roughly 30% to 40% of the potential high-tier workforce. - The systemic absence of women in senior leadership leads female professionals to voluntarily opt out due to a lack of visible advancement paths. - Men disproportionately self-advocate and ask for stretch assignments, requiring leaders to intentionally seek out qualified candidates who do not self-promote.

---

[14:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=870s) Navigating Interview Red Flags, Culture Fit, and Personal Networks

For early-career and junior women navigating the AI interview circuit and encountering dismissive attitudes, Georgiadis advises reframing aggressive or outdated workplace expectations. Recalling a mentee who was told she would need to "wear a beeper" to prove her dedication, Georgiadis suggests turning the question around directly:

> "Well, I'm glad you talked about wearing the beeper. I probably outwork most of my male colleagues today. Happy to do a beeper if it's really necessary. What I really care about is impact and doing the work that I want to do."

She encourages candidates to test whether an interviewer's demand is merely a test of work ethic or indicative of a genuinely toxic environment. If an organization's values do not align with their own, candidates should feel empowered to self-select out. The most dependable way to identify organizations that actively nurture and develop talent is through extensive personal networking—putting "the miles on the feet, on the phone, on the video conferences" with trusted industry peers who know how those companies operate from the inside.

---

[16:22](https://www.youtube.com/watch?v=wDLGYWBalvs&t=982s) Bridging the AI Confidence Gap and Executive Upskilling

Turning to legacy industries integrating artificial intelligence, Georgiadis addresses the persistent confidence gap between men and women. Citing recent survey figures, she notes that approximately 80% of women in professional roles feel they receive insufficient education around AI tools. Across the broader population, only about 30% of women report feeling comfortable with AI compared to 71% of men. With a laugh, she shares that she had stress-tested the latest large language models before taking the stage, testing ChatGPT against Google Gemini and giving the edge to ChatGPT despite being an "ex-Googler."

To successfully integrate AI into legacy enterprises, Georgiadis argues that C-suite leaders must personally commit two to three hours every week to direct AI immersion—reading industry-specific research, consulting peers, and hands-on stress-testing the tools with difficult questions. Finally, she highlights the vital role of "change integrators," who combine deep domain expertise with empathy and forward-looking vision to guide staff through technological transitions, a leadership strength where women frequently excel.

- Roughly 80% of professional women report that they do not receive adequate workplace education in AI. - A significant confidence divide exists across the general population, where only ~30% of women feel comfortable with AI compared to 71% of men. - C-suite executives should dedicate 2 to 3 hours per week to hands-on testing, reading, and peer discussions to effectively steer enterprise AI transformation.

[18:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=1105s) Overcoming Cultural Resistance and Closing Remarks

Addressing the organizational challenges that accompany technological transformation, Margo Georgiadis reflects on what is truly possible within a modern company and how leaders must help their teams process and digest these rapid shifts. She emphasizes the necessity of fundamentally resetting how work gets done and how success is measured, underscoring that employees must feel genuinely included in the transition. Without that sense of inclusion, Georgiadis warns, overcoming deep-seated cultural resistance becomes nearly impossible. She observes a tremendous amount of active work currently taking place across these organizational fronts and expresses strong conviction that women are uniquely positioned to play a leading role in championing this cultural change.

Emma Hinchliffe then brings the fireside chat to a close, expressing her gratitude to Georgiadis for the insightful conversation and thanking the audience for attending.

- **Organizational Evolution**: Transforming enterprise workflows requires resetting baseline metrics of success alongside helping employees actively embrace emerging technological possibilities. - **Cultural Inclusion**: Ensuring broad participation and inclusion is vital for dismantling internal friction and organizational pushback during AI adoption. - **Women in Leadership**: Female leaders are viewed as pivotal drivers in shaping corporate culture and steering inclusive organizational transformation in the AI era.

## Notable quotes > [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) just a war — Margo Georgiadis

> [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) probably the biggest $150 billion company you've never heard of — Margo Georgiadis

> [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) And you're going to have to be on a beeper 24 hours a day, seven days a week — Hiring Manager

> [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) bro culture sexism — Margo Georgiadis

> [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) slinging beers at the bar afterwards — Margo Georgiadis

> [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) Remind yourself that if you don't put yourself in the path of opportunity and let people know what you're interested in and ask them, 'Do you think I'm qualified? And if I'm not, what do you think I need to do?' If you don't make that effort, you don't remind them that you want that. — Margo Georgiadis

> [14:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=870s) Well, I'm glad you talked about wearing the beeper. I probably outwork most of my male colleagues today. Happy to do a beeper if it's really necessary. What I really care about is impact and doing the work that I want to do. — Margo Georgiadis

> [14:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=870s) the miles on the feet, on the phone, on the video conferences — Margo Georgiadis

## Mentioned Ai4, Ancestry, AppLovin, ChatGPT, Emma Hinchliffe, Fidji Simo, Fortune, Google, Google Gemini, Grok, Margo Georgiadis, Mattel, McDonald's, Mira Murati, Montai Therapeutics, Tesla, Thinking Machines Lab

## Video description Emma Hinchliffe, Editor, Fortune Most Powerful Women, Fortune Margo Georgiadis, Co-Founder & CEO, Montai Therapeutics

Abstracts

A candid fireside chat with leading women in AI, exploring their journeys, insights, and impact on the future of the field.

[Source](https://www.youtube.com/watch?v=wDLGYWBalvs)

中文简报

Montai CEO Georgiadis警告：AI初创公司或流失40%顶尖人才仅依赖高额薪酬而非包容且高度自主文化的AI企业，将失去高达40%的顶级技术人才。 Margo Georgiadis探讨了Montai Therapeutics利用AI应对慢性疾病的使命，并阐释了为何在AI时代，企业的护城河取决于分发渠道与工程文化，而不仅是薪酬。她详细分析了女性技术人才面临的结构性障碍以及普遍存在的“兄弟文化”（bro culture），强调算法偏见和代表性不足会直接危害产品功效。Georgiadis提出了推进包容性招聘、弥合女性自我倡导差距的实用策略，并呼吁管理层亲身投入AI技能提升，以引领组织变革。 Montai Therapeutics 利用 AI 重构慢性病小分子药物研发，致力于填补仅有不到 10% 确诊患者能获得突破性疗法的医疗缺口，目前正推进其首批 4 款候选药物进入临床阶段。 (0:04) 随着基础大模型向垂类软件渗透，AI 初创企业的护城河取决于分发能力以及成为核心记录应用（application of record）；吸引顶尖技术人才更需依托授权与即时交付生产代码的工程师文化，而非单纯拼抢高额薪酬。 (1:15) 要求 24/7 待命等形式主义测试与‘兄弟会文化’劝退了业内近 70% 的女性；科技公司应重写注重互补能力的招聘需求，并通过数据审计筛查面试中的系统性偏见。 (4:11) 历史上仅 1% 的科研针对女性特有疾病，此类数据缺失要求企业必须将多元包容列为核心业务 OKR，以规避算法偏见并化解生命科学等高专业度领域内部对 AI 的文化抵触。 (9:25)

Reading summary

Margo Georgiadis argues that long-term defensibility in artificial intelligence relies on application-layer distribution and human-centered engineering cultures, warning that excluding diverse talent and perspectives introduces critical algorithmic flaws and alienates top performers. Application-Layer Distribution and Talent Culture as Moats (1:15) Georgiadis describes the intensifying competition between foundation model builders and vertical software as a market war that is pushing founders up the tech stack. In her assessment, defensibility will not stem from models alone, but from establishing entrenched distribution as the primary "application of record"—particularly in high-stakes sectors like healthcare where users are sensitive to incremental differences. Simultaneously, she contends that startups cannot win the engineering talent race through outsized compensation packages alone. Highlighting AppLovin, she notes that providing early-career engineers and interns with autonomy to ship production code directly creates an authentic culture of impact. This environment has successfully attracted top talent away from competitors such as Grok and Tesla, showing that high performers—particularly women—weigh belonging and meaningful work over pure compensation. Systemic Bias and Representation in Product Architecture (9:25) Because artificial intelligence fundamentally restructures economic and human outcomes, Georgiadis warns that homogenous design teams risk cementing historical biases into modern infrastructure. She points out that historically, only 1% of scientific research has targeted diseases specific to women, while automated hiring algorithms routinely amplify demographic imbalances by cloning past employee résumés instead of assessing holistic capabilities. At Montai Therapeutics, addressing healthcare disparities is central to applying AI to small-molecule drug discovery, given that fewer than 10% of all diagnosed patients and 30% to 50% of severely diagnosed patients currently have access to transformative therapies. Drawing from her tenure as CEO of Ancestry, Georgiadis emphasizes that inclusive product design is equally crucial across consumer and enterprise environments. Product teams building family history and health tools had to intentionally reflect diverse immigrant and ethnic backgrounds in user interfaces and models. In B2B fields like life sciences, adoption often hinges on addressing the anxieties of deeply specialized experts—such as researchers focused on single cell types—who resist multifactorial AI tools out of fear their expertise will become obsolete. Reforming Tech Culture and Hiring Pipelines (4:11) Pervasive workplace attitudes continue to deter female technical leaders, with Georgiadis noting that roughly 70% of women in technology report experiencing exclusionary "bro culture," defined by pressure to work excessive hours or socialize in male-dominated settings. She cites an incident where a premier engineering candidate withdrew after an interviewer demanded 24/7 beeper availability, illustrating how aggressive framing alienates elite candidates. For early-career women confronting such red flags, she recommends turning the question around to evaluate whether the demand reflects a legitimate business necessity or a toxic environment, while using professional networks to identify supportive workplaces. To counter structural bias, Georgiadis urges organizations to overhaul talent acquisition by rewriting job descriptions to prioritize complementary skills rather than homogenous profiles. She also advocates using AI tools to audit hiring pipelines and analyzing interview scoring trends. Citing her past leadership at Google, she recalls that data audits uncovered specific interviewers who systematically failed diverse candidates and candidates of color regardless of their objective qualifications. Addressing Self-Advocacy and the AI Literacy Divide (11:57) Georgiadis frames diversity as an unavoidable business calculation, cautioning male-dominated leadership teams that ignoring inclusion means willingly forfeiting 30% to 40% of elite technical talent. When women see no visible advancement paths into executive ranks, they routinely opt out of organizations. She also identifies a behavioral divide where male employees disproportionately lobby for promotions and stretch projects, whereas qualified women ask far less frequently. She urges leaders to proactively identify overlooked talent while encouraging women to place themselves directly in the path of opportunity by seeking clear feedback on qualifications. At the enterprise level, adopting AI requires closing a pronounced confidence gap: survey data indicates that 80% of professional women feel they lack sufficient AI education, and only 30% of women report feeling comfortable with AI compared to 71% of men. Georgiadis recommends that C-suite executives personally dedicate two to three hours weekly to hands-on testing, reading research, and consulting peers. She highlights the need for empathetic "change integrators" to reset organizational metrics and guide workforces through cultural disruption. While Georgiadis asserts that human-centered leadership and diverse teams are essential to deploying AI safely and effectively, she acknowledges that overcoming entrenched cultural resistance among specialized workforces and legacy leadership remains an ongoing organizational challenge.

精炼阅读摘要

Montai Therapeutics 首席执行官 Margo Georgiadis 指出，AI 领域的长远护城河在于分发能力与成为不可替代的“记录级应用”，而非单纯的模型规模；企业若忽视包容性文化与多元视角，不仅会加剧算法偏见并引发落地阻力，还将主动错失 30% 至 40% 的顶尖技术人才。软件应用层护城河与工程文化重构 (1:15) Georgiadis 认为，随着基础模型层向纵向垂直工具扩张，AI 初创团队正加速转向应用层。在医疗健康等高风险、高敏感度领域，构筑长期防御能力的核心在于建立分发壁垒，成为用户深度依赖的“记录级应用”（application of record）。以其领导的 Montai Therapeutics 为例，该公司运用 AI 从第一性原理切入慢性病小分子药物研发，旨在弥补现有诊断患者中仅不到 10%、重症患者中仅 30% 至 50% 获得变革性疗法的医疗缺口，目前已准备推进首批 4 款候选药物进入临床阶段。面对激烈的工程人才争夺，Georgiadis 指出单纯比拼薪酬方案已难以维系优势，关键在于建立以信任和自主权为核心的工程文化。她以估值约 1500 亿美元的 AppLovin 为例，说明让年轻工程师乃至实习生直接参与生产环境代码发布、提供即时业务影响力的机制，不仅成功从 Grok 和 Tesla 等机构吸引顶尖人才，也更契合女性等高绩效技术群体对归属感与实质性贡献的追求。破除“兄弟文化”与重构招聘流程 (4:11) 针对技术团队拒绝巨额薪酬包的行业现象，Georgiadis 强调工作文化对人才留存起着决定性作用。她指出科技领域约 70% 的女性经历过排他性的“兄弟文化”（bro culture），并分享了一位女性工程高管在应聘头部基础模型公司时，因面试官提出“必须 7x24 小时随叫随到”的苛刻要求而当场终止面试的案例。Georgiadis 警示管理层，忽视包容性意味着企业在人才招募中主动放弃了 30% 至 40% 的优质劳动力；而高管层缺乏女性代表，则会导致女性员工因看不见晋升通道而系统性退出。为消除结构性障碍，Georgiadis 呼吁企业利用数据与 AI 工具重构招聘体系。她回顾在 Google 的任职经历，指出当时的数据审计揭示出部分面试官无论候选人资历如何，均系统性给予多元化背景及有色人种候选人不及格评分。她建议企业重新撰写岗位要求以强调技能互补性，利用算法审计排除面试偏差，并剔除带有偏见的面试官。算法偏见风险与人本技术落地 (9:25) Georgiadis 强调，AI 正在深刻重构全球经济与工作范式，若缺乏多元群体参与架构设计，技术系统将重蹈历史偏见的覆辙。她列举数据指出，历史上仅有 1% 的科学研究聚焦于女性特有疾病；而在招聘领域，若直接沿用历史简历训练筛选算法，工具只会机械复制既有员工画像。在担任 Ancestry 首席执行官期间，她要求团队在开发 AI 家族历史与健康产品时，必须融入多元族裔与移民视角的考量，以确保产品界面和模型输出具备广泛的适应性与准确性。在 B2B 企业端，技术落地同样面临深刻的文化与心理阻力。例如在生命科学领域，专注特定细胞类型数十年的专家往往因担心专业积累被取代，而对多因子 AI 整合工具产生抵触。Georgiadis 指出，企业必须将多元包容设为核心业务目标（OKR），通过流程共建与技能再培训来缓解员工对技术颠覆的焦虑。弥合信心差距与高管 AI 领导力 (16:22) 围绕企业 AI 转型，Georgiadis 披露了显著的性别信心分歧：约 80% 的职场女性反映未能在工作中获得充分的 AI 教育培训；而在普通公众中，仅约 30% 的女性表示对使用 AI 感到自如，相较之下男性的比例达到 71%。此外在职业发展上，男性通常更频繁地主动向管理层要求升职或争取核心项目，而女性较少主动倡导自己。她一方面呼吁管理者有意识地发掘不主动推销自己的高潜力员工，另一方面建议女性主动表达职业兴趣并站上关键通道。对于推进智能化转型的传统企业高管，Georgiadis 建议最高管理层（C-suite）每周固定投入 2 至 3 小时，亲自对主流大语言模型进行压力测试、研读垂直行业报告并与同行交流。她特别指出，兼具行业深厚积累与人本同理心的“变革整合者”（change integrators）是引导员工平稳度过转型期的中枢，而女性领导者在协调这种技术与人性的平衡中具备独特优势。 Georgiadis 总结强调，AI 驱动的组织转型关键在于从根本上重塑工作流程与成功衡量标准，必须确保全员在心理与机制上融入变革过程，否则技术应用将难以逾越深层的组织与文化阻力。

Detailed notes — zh

## Background 本内容由 YouTube 频道 Ai4 于 2025 年 8 月 25 日发布。页面描述显示，这是一场由《财富》（Fortune）杂志“最具影响力女性”编辑 Emma Hinchliffe 与 Montai Therapeutics 联合创始人兼 CEO Margo Georgiadis 展开的炉边对话，深入探讨了女性在 AI 领域的职业历程、前沿见解及其对未来行业发展的深远影响。 ## TL;DR 对话围绕 AI 在生物医药创新、企业生存壁垒与组织人才文化等维度的深度融合展开。Margo Georgiadis 分享了 Montai 利用 AI 研发小分子药物攻克慢性病的愿景，并指出在残酷的“AI 军备竞赛”中，应用层的分发壁垒与赋能型工程文化才是真正的胜负手。她深入剖析了科技招聘中普遍存在的“兄弟会文化”与数据偏见，呼吁建立更具包容性的技术架构与招聘审计机制。同时，她鼓励女性主动争取晋升跑道、打破 AI 信心鸿沟，并强调女性领导者凭借同理心与整合力，将在推动组织 AI 文化变革中扮演至关重要的角色。 ## Atmosphere & style 整场炉边对话节奏紧凑、氛围热烈而真诚。Emma 的提问精准切中行业痛点，而 Margo 的回答兼具前瞻战略视野与细腻的人性关怀，从跨国企业管理经验到实战面试案例信手拈来。现场充满力量感与启发性，既有对技术残酷竞争的理性剖析，也有对女性职场破局的热情鼓舞，给听众带来极具穿透力与共鸣感的高管思维洗礼。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=wDLGYWBalvs&t=4s) Montai Therapeutics 致力于利用 AI 改造小分子药物研发流程，以解决慢性病领域高达 30%–50% 重症患者无法获取突破性疗法的医疗鸿沟。 - [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) 在基础大模型不断侵蚀应用层的市场竞争中，AI 原生企业的立足核心在于建立强大的分发渠道（Distribution）并占领行业阵地，成为不可替代的“记录级核心应用”（Application of Record）。 - [02:00](https://www.youtube.com/watch?v=wDLGYWBalvs&t=120s) 单纯依赖高薪的人才军备竞赛不可持续，AI 企业需通过扁平化管理、高度信任以及让工程师快速开发并上线真实产品，构建差异化的赋能型工程文化。 - [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) 科技与 AI 行业中约 70% 的女性仍面临“兄弟会文化”与隐性偏见；以“24/7 随叫随到”等形式主义工作狂指标考核员工，容易错失注重实际业务影响力的顶尖女性高管。 - [05:10](https://www.youtube.com/watch?v=wDLGYWBalvs&t=310s) 企业应当重塑包容性岗位描述，并借助 AI 反向审计招聘流程中面试官的评分偏好，精准排查并纠正系统性的性别与族裔筛选偏见。 - [06:38](https://www.youtube.com/watch?v=wDLGYWBalvs&t=398s) AI 底层架构设计必须将纯技术指标与人类福祉深度绑定；决策层若缺乏多元视角把关，算法偏见与误判将导致技术难以被社会大众信任。 - [09:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=565s) 历史上仅有 1% 的医学研究针对女性特有疾病，数据缺失会导致算法同质化；企业应将消除偏见与包容性设计列入核心业务 OKR。 - [10:40](https://www.youtube.com/watch?v=wDLGYWBalvs&t=640s) 在生命科学等高度专业化的 B2B 领域，传统专家面对 AI 集成化趋势容易产生职业危机感，企业技术落地需要同理心与文化层面的正向引导。 - [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) 将“AI 军备竞赛”作为忽视多元化的借口是短视的，这等于主动放弃劳动力市场中 30%–40% 的顶尖人才；高管层缺乏女性榜样将导致女性员工因看不到路径而流失。 - [13:10](https://www.youtube.com/watch?v=wDLGYWBalvs&t=790s) 女性在职场中应主动站到机会跑道上表达抱负，管理者也需敏锐意识到性别与文化差异，主动给予认可并建立公平的晋升评估机制。 - [14:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=870s) 面对面试或职场中的不合理偏见测试，女性应自信聚焦于业务影响力和实际价值创造，并果断进行“自我筛选”，离开有毒文化并借助人脉寻找契合自身价值观的团队。 - [16:22](https://www.youtube.com/watch?v=wDLGYWBalvs&t=982s) 行业内存在显著的“AI 信心鸿沟”，普通人群中仅约 30% 女性对使用 AI 感到自信（男性为 71%），企业亟需普及系统化 AI 教育培训以打破技能壁垒。 - [17:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=1050s) C-Suite 高管每周应固定投入 2 到 3 小时深度实操与极限压力测试 AI 工具，同时企业应大力提拔兼具业务专长与共情力的女性作为“变革整合者”。 - [18:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=1105s) 组织推进 AI 转型的最大障碍往往来自文化层面的抵触，领导者必须重塑工作流与成功衡量标准，营造包容感以确保员工在技术重构中拥有切实的归属感。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=wDLGYWBalvs&t=4s) 开场介绍与 Montai Therapeutics 的创立初衷：用 AI 攻克慢性病小分子药物论坛在轻松而热烈的氛围中开场。《财富》（Fortune）杂志“最具影响力女性”（Most Powerful Women）编辑 Emma Hinchliffe 首先欢迎全场观众，看到座无虚席的会议厅，她感叹在 Ai4 峰会上能汇聚如此多卓越的女性领军者令人振奋。Emma 随后介绍了本场嘉宾 Margo Georgiadis 的重磅履历：她曾担任知名家谱与基因检测平台 Ancestry 的 CEO，并在麦当劳（McDonald's）、美泰（Mattel）以及谷歌（Google）等全球巨头担任高管或董事会成员，如今她作为联合创始人兼 CEO 正在全力打造生物科技公司 Montai Therapeutics。被问及创立 Montai 的初衷与业务愿景时，Margo 动情地阐述了团队对利用小分子药物治疗慢性病的执着追求。她指出，生命科学领域尽管涌现了大量创新疗法，但绝大多数根本无法惠及广泛的患者群体。数据显示，在确诊病情最严重的患者中，有 30% 至 50% 无法获得能够改变生命的特效药物；若放眼全部确诊患者，能够用上创新疗法的比例甚至不足 10%。正是面对这一巨大的医疗鸿沟，Montai 团队萌生了利用 AI 重构药物发现全流程的构想，旨在大幅降低研发门槛并提高药物可及性。目前 Montai 正推进其首批 4 款创新疗法迈向临床试验阶段。Margo 坦言，自己并非生命科学专业出身，跨界进入这个门槛极高的行业并从零学起，是一段极其艰难但收获颇丰的历程。 - **慢性病医疗痛点**：重症确诊患者中有 30%–50% 无法获得突破性治疗药物；在全部确诊患者群体中，获得前沿治疗的比例不足 10%。 - **研发管线进展**：Montai 借助 AI 驱动的分子发现与重塑架构，目前已有首批 4 款核心在研疗法正在全力推进至临床阶段。 --- ### [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) AI 原生企业的生存之战：应用层护城河与超越高薪的人才文化探讨当前科技产业格局时，Emma 提出 Margo 同时拥有观察原生 AI 初创企业与传统跨国企业整合 AI 进程的独特视角，并请她剖析 AI 原生企业（AI-native startups）当前面临的最大瓶颈。Margo 直言，无论是基础大模型公司还是垂类 AI 应用软件企业，当前整个市场都处于极其残酷的“血腥战争”（"it's just a war right out there"）中。随着基础模型能力不断演进，越来越多的团队正被迫转向应用层，大家都在焦虑地思索：究竟什么样的应用软件能够成为具备抗风险能力的常青业务，又有哪些应用终将被日益强大的基础模型直接吞噬和取代。 Margo 强调，应用层能否立足最终取决于**分发能力（Distribution）**与**行业阵地占领（Toll hold）**。一旦某个产品成为特定行业不可替代的“记录级核心应用”（application of record），竞争对手和通用模型就极难将其撬动；在医疗健康等对细微差异与精度极度敏感的严苛领域，这种壁垒尤为牢固。然而，原生 AI 企业面临的另一大死穴是顶尖人才争夺战（War for talent）。在全行业盲目哄抬天价薪酬包的当下，Margo 指出单纯比拼金钱已难以为继，文化建设才是真正的胜负手。她以自己担任顾问及独立董事的移动广告平台 AppLovin 为例——这家市值高达 1500 亿美元的“隐形巨头”通过重构广告投放引擎实现了惊人增长，但更令她钦佩的是其工程团队倡导的文化：高度信任年轻工程师，提供扁平化架构，让他们能够快速开发并上线真实产品（"make and ship things quickly"）。在近期举办的实习生展示会上，哪怕是实习生也在交付真正影响业务的核心项目，而非端茶倒水打杂。这种让员工深感自身价值被认可的环境，使他们能从 xAI（Grok）、特斯拉（Tesla）等头部大厂手中持续吸引顶级工程人才。 ```mermaid flowchart TD accTitle: AI 原生企业生存与人才竞争双轮驱动模型 subgraph MarketMoat["市场生存壁垒构建"] A["基础模型快速演进与侵蚀风险"] --> B["转向垂类应用层 (Application Layer)"] B --> C["建立核心分发渠道 (Distribution)"] C --> D["成为行业核心记录应用 (Application of Record)"] end subgraph CultureStrategy["差异化人才争夺策略"] E["恶性薪酬军备竞赛 (Pay War)"] -.->|不可持续| F["构建赋能型工程文化"] F --> G["扁平化管理与充分授权"] G --> H["快速构建并上线产品 (Ship Quickly)"] H --> I["提供真实业务影响力与归属感"] end D --> J["长期商业抗风险壁垒"] I --> J ``` --- ### [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) 招聘中的隐性偏见：如何击碎“兄弟会文化”以吸纳顶尖女性领袖对话延伸至女性技术人才在 AI 领域的留存与发展。Emma 提及前 OpenAI 首席技术官 Mira Murati 以及近期成立的 Thinking Machines Lab 团队面临其他大厂开出数亿美元天价挖角却依然坚守的案例，询问企业文化与价值观在其中起到了何种关键作用。Margo 借此分享了自己在辅导众多高阶女性技术管理者时听到的真实且令人震惊的面试遭遇：一位顶尖的女性产品与工程负责人近期受邀面试某头部大模型基础公司，面试过程中对方男性面试官反复强势强调“你必须每周 7 天、每天 24 小时随身带着寻呼机待命”（"you're going to have to be on a beeper 24 hours a day, seven days a week"）。该面试官自以为这是在考察候选人的“敬业与投入度”，但这种傲慢而脱节的沟通方式极具攻击性，导致该女性候选人当场终止面试并断然离席。 Margo 一针见血地指出，这位女性候选人日常的工作拼劲超越身边的绝大多数男性同行，面试官若与她探讨产品与技术影响力，本可达成极佳的合作，但他充满说教且缺乏同理心的态度彻底断送了招募顶级人才的机会。这一案例折射出行业普遍现象：调查显示高达 70% 的女性在科技与 AI 行业中能深切感受到“兄弟会文化”（bro culture）与隐性性别歧视——如果女性无法接受无休止的非理性加班，或者在下班后不愿混迹于酒吧跟男同事推杯换盏，就会被边缘化并视为“不够融入”。Margo 呼吁领导者必须系统性重塑招聘机制： 1. **重写包容性岗位描述**：消除措辞中的男性化偏见，重新定义团队真正需要的多元化技能矩阵。 2. **运用 AI 工具反向审计招聘漏斗**：分析候选人在各个面试轮次的通过率，追溯面试官的评分偏好。Margo 回忆在谷歌推进此类分析时曾发现，某些特定面试官的名下，少数族裔及女性候选人的通过率几乎为零；精准识别出这些系统性偏差后，企业得以及时调整面试官阵容，真正实现技术向善。 - **行业文化现状**：约 70% 的女性科技从业者表示在职场中遭受过“兄弟会文化”或隐性性别排斥的困扰。 - **面试考核误区**：以“24/7 随叫随到”等形式主义工作狂标准衡量承诺度，严重阻碍了看重实际业务影响力的成熟高管人才引入。 --- ### [06:38](https://www.youtube.com/watch?v=wDLGYWBalvs&t=398s) 影响力的多维诠释：为何构建具备包容性的人类 AI 架构至关重要针对 Emma 提出“在非生命科学等天然具有崇高使命感的行业中，如何向看重‘影响力’的女性候选人抛出橄榄枝”的问题，Margo 再次以广告技术公司 AppLovin 举例说明。虽然广告分发业务看似缺乏直接拯救生命的崇高光环，但其底层是一个算力庞大、类似 DeepSeek 的尖端 AI 引擎。在此类硬核技术企业中，向女性传递影响力的核心在于证明组织对创新的尊重与人性关怀：“在这里，你的创意会被真正采纳，组织架构极度扁平，管理层发自内心地支持你冒险创新并视你为独特的人才。”她甚至分享了一个动人细节：AppLovin 负责海外工程团队的负责人曾专门为员工设立为期 4 天的带薪“孝亲假期”，要求员工必须陪伴父母度假，这种尊重本土文化、充满人性温度的关怀对于凝聚多元人才具有极大的感召力。最后，Margo 站在行业宏观视角深刻阐述了女性在 AI 领导层中不可或缺的战略意义。面对业界常有“当前处于 AI 军备竞赛期，追求多元化是否拖慢速度”的质疑，Margo 坚定地反驳：当前整个 AI 产业正处于重塑人类生产方式、产业格局与社会生活的历史拐点；AI 算法与产品的底层架构设计将深刻作用于每一个普通人的命运，因此架构设计本身必须从一开始就具备广泛的包容性。她特别称赞了 Mira Murati 对开源与开放生态的坚持，并对 Instacart 前 CEO Fidji Simo 加入 OpenAI 董事会及相关架构探索表示赞赏——强调设计 AI 时绝不能只盯着纯技术指标（technical outcomes），必须将“技术结果”与“人类福祉结果”（human outcomes）深度绑定。Margo 告诫全场，回顾过去 5 到 20 年的 AI 发展史，警务预测系统中的算法偏见以及医疗诊断模型在特定人群上的严重失误屡见不鲜；若缺乏多元视角在核心决策层的把关，技术不仅无法被社会大众信任与采纳，更可能重蹈覆辙引发灾难性后果。 ### [09:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=565s) 消除数据偏见：从 Ancestry 经验到 B2B/B2C 的多元包容实践在医疗健康与算法模型中，偏见往往源自基础数据的缺失与失真。Margo Georgiadis 指出，过去仅有 1% 的医学研究真正聚焦于女性特有疾病，而缺乏正规医疗保障的多元化群体在临床数据中的代表性严重不足，这导致模型不可避免地继承了数据固有的系统性偏见。招聘领域同样如此，若筛选简历的 AI 机器人仅以现有员工画像为训练基准，系统就会不断复制同质化人才，忽视更广泛的技能背景。对此，优秀的企业不再将 AI 偏见仅仅视作技术挑战，而是将其转变为业务战略重点，通过设立专门的 OKR 来确保产品开发与评审全流程都有多元化视角的介入。 ```mermaid flowchart TD accTitle: 数据偏见识别与多元化产品流程 subgraph S1["数据与算法挑战"] A["历史偏见数据 (医疗/招聘/临床样本)"] --> B["同质化算法输出 (固化现有结构)"] end subgraph S2["战略解决路径"] B --> C["设定业务 OKR (将包容性列为必选项)"] C --> D["多元团队参与产品全流程 (UI/UX、市场与用户反馈)"] D --> E["赋能 B2B/B2C 落地 (化解文化与技能转型阻力)"] end ``` 这种包容性设计在消费端（B2C）与企业端（B2B）同样至关重要。Margo 分享了她在 Ancestry 担任高管时的经历：团队在开发 AI 健康新产品和家族历史溯源体验时发现，不同族裔进入美国的历史背景迥异，用户对同一功能的感受大相径庭。因此，每一次产品迭代都必须由多元化团队深入审视产品逻辑、用户界面（UI）体验与市场推广策略，确保对所有群体的尊重。而在 B2B 领域（特别是生命科学行业），许多专家在特定细胞类型或狭窄领域深耕了 20 年，高度专业化的背景容易让他们在面对多因素整合的 AI 工具时产生巨大的防御心理，担心自身工作被取代。因此，无论 B2B 还是 B2C，企业都需要在技术整合中兼顾同理心与文化引导。 - **数据偏差现实**：历史上仅 1% 的医学研究针对女性特有疾病，少数族裔在临床健康数据中存在显著缺失。 - **算法同质化陷阱**：招聘 AI 若仅按现有成功样本训练，将不断自我强化，过滤掉具备跨领域潜力的候选人。 - **专业人才转型阻力**：生命科学等领域的顶尖专家容易因 AI 集成化趋势而产生职业危机感，需要企业在变革中建立引导与信任。 --- ### [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) 应对“AI 军备竞赛”论：主动发声与打破晋升盲区针对当前行业内普遍弥漫的“AI 处于战争时刻、无暇顾及多元化”的论调，Margo 展现出清晰坚定的反驳逻辑。在面对持此类观点的企业决策层时，她直言不讳地提醒对方：“你们真的愿意直接放弃劳动力市场中 30% 到 40% 最顶尖、最专注的人才吗？”她指出，当高管层全部由男性构成且仅依照自身舒适圈行事时，往往难以察觉体制性障碍。如果没有女性身处最高管理层，中基层女性往往会因为看不到合理的发展路径而主动选择退出（opt out）。 Margo 进一步剖析了职场晋升中的行为差异：在实际管理中，男性员工往往会频繁主动敲开她的办公室大门，自信满满地陈述自己完全有资格胜任下一个重要职位或承担全新项目；相比之下，女性员工主动争取的比例要低得多。领导层若没有意识到这种文化和性格差异，就会在潜意识中把每一次升职和关键任务分配给那些最主动推销自己的男性。 - **对女性的行动建议**：必须主动站到机会的跑道上（"put yourself in the path of opportunity"），清晰告知管理者自己的抱负，并主动询问：“你认为我合格吗？如果不合格，我需要提升哪些方面？” - **对管理者的反思要求**：管理层必须敏锐觉察到不同文化背景及女性群体倾向于不主动开口的特质，做出微小但关键的领导力调整。 - **榜样效应**：高管层缺乏女性代表会导致系统性的人才流失，领导者只需付出极少的时间去主动表达认可并规划职业路径，即可留住核心人才。 --- ### [14:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=870s) 初入职场女性的破局之道：识别企业文化与网络筛选对于身处面试或职业初期、对 AI 充满热情却因遭遇偏见而产生动摇的年轻女性，Margo 传授了极具实战价值的心态与应对策略。她以此前辅导的一位女高管为例——对方在面试中曾被男性考官刁难是否愿意“随叫随到佩戴寻呼机（beeper）”。Margo 建议在遇到此类试探时，不必内耗或将其个人化，而是应当自信地将问题抛回给对方：“我很高兴你提到随时待命。事实上，我今天的工作产出已经超过了大多数男性同行。如果真正有必要，我非常乐意待命；但比起形式，我更看重的是实际影响力与工作价值。” ```mermaid flowchart LR accTitle: 女性在职场面试与环境筛选中的应对路径 A["遭遇职场偏见/不合理测试"] --> B["自信反问与重构对话 ('聚焦业务影响力而非无谓内卷')"] B --> C{"该组织是否契合自身价值观？"} C -- 否 --> D["果断放弃有毒环境 (Self-select out)"] C -- 是 --> E["借助人脉深入调研 (了解人才培养与晋升机制)"] ``` Margo 坦诚地告诫求职者，并非所有组织都适合每个人，女性应当果断自我筛选（self-select），选择与自身价值观契合的企业。她强调，寻找真正健康且重视人才成长的平台，不能仅依赖公开招聘信息，而是必须下足功夫（"put the miles on the feet"），充分调动人脉网络，通过电话、视频与深入对话向熟知目标机构运作机制的同行求证，从而找到能够让自身才华得以施展的成长土壤。 - **化解挑衅技巧**：厘清对方是在测试你的敬业度还是在兜售形式主义文化；如果是前者，积极证明决心；如果是后者，及时抽身。 - **主动自我筛选**：不要强求融入缺乏包容性的环境，寻找认同自身价值的团队才是长远发展的基石。 - **深耕人脉调研**：通过熟知内情的业界同行，透彻了解目标公司如何对待人才、培养人才与评估价值。 --- ### [16:22](https://www.youtube.com/watch?v=wDLGYWBalvs&t=982s) 弥合 AI 信心鸿沟：高管深耕学习与“变革整合者”的力量在谈及传统行业向 AI 转型的过程中，Margo 指出职场女性目前面临着严峻的“AI 信心鸿沟”。她在参会前查阅并对比了 ChatGPT（GPT-5）与 Gemini 的相关数据（她幽默地表示，尽管自己是前谷歌人 ex-Googler，但在这项特定测试中不得不承认 GPT-5 略胜一筹）。数据显示，专业岗位中高达 80% 的女性感到自己未能获得充分的 AI 教育培训；而在普通人群中，仅有约 30% 的女性对理解和应用 AI 感到自信，而男性的这一比例高达 71%。这种认知与教育的落差，极大地限制了多元化人才进入 AI 核心舞台。 | 评估维度 | 调研与行业现状 | 应对与破局策略 | |:--- |:--- |:--- | | **AI 认知与信心比例** | 约 30% 女性感到适应 vs 71% 男性感到适应 | 强化企业级 AI 普及培训，打破工具接触壁垒 | | **高管学习投入** | 许多传统企业 C-Suite 缺乏下场实操的决心 | 每周固定投入 2 至 3 小时：阅读前沿、同行交流、极限压力测试工具 | | **转型领导角色** | 变革常因缺乏同理心与业务理解而受阻 | 提拔具备深厚业务积累与共情力的“变革整合者”（女性具备显著优势） | 为了从根本上消除这一差距并在企业内部成功落地 AI，Margo 提出了两条核心主张：首先，C-Suite 级别的高管必须躬身入局，每周至少投入 2 到 3 个小时用于 AI 专属学习——不仅要阅读行业相关资料、与同行深入交流，更要亲自上手使用工具并提出最严苛的问题进行压力测试；没有最高管理层的全情投入，AI 转型注定流于形式。其次，企业亟需重用“变革整合者（change integrators）”，而女性在这一角色中往往展现出独特优势——她们不仅具备深厚的业务领域专长，更拥有敏锐的前瞻洞察力和体察员工情绪的同理心，能够平稳引领组织跨越技术重构的阵痛。 ### [18:25](https://www.youtube.com/watch?v=wDLGYWBalvs&t=1105s) 组织变革中的包容性与女性在重塑工作模式中的核心力量在探讨组织面对技术浪潮的应对方式时，Margo Georgiadis 强调了深刻理解“组织内部究竟有何新可能”的关键意义。她指出，领导者的核心任务在于帮助员工真正理解并消化这些前沿技术带来的冲击，进而重置团队的工作方式与衡量成功的标准。这一过程的关键在于营造包容感，让每个人都觉得自己在变革中占有一席之地；因为一旦员工感到被边缘化或被排除在外，组织将面临巨大的阻力，“要克服那种文化层面的抵触情绪就会变得异常艰难”。Margo 观察到，目前许多企业已经开始在文化适应与机制重塑方面投入大量精力，而她坚信女性领导者凭借其独特的同理心、沟通力与整合视角，完全可以在引领这场深刻变革中发挥举足轻重的作用。随着对话进入尾声，主持人 Emma Hinchliffe 向 Margo 致以诚挚的谢意，称赞这是一场精彩绝伦且启发深刻的炉边对话，并对现场所有参会观众的聆听与支持表示感谢。Margo 亦向主持人与全场观众致意道谢，为整场关于女性 AI 领导力、生物医药创新与未来组织演进的讨论画上了圆满的句号。 - **组织适应与文化接纳**：推动 AI 等新技术落地的最大阻碍往往不在于技术本身，而在于文化层面的阻力；领导层必须重塑工作流与成功评估体系，确保团队成员拥有切实的参与感与归属感。 - **女性在变革中的优势**：女性管理者在协调组织内部关系、化解文化抵触以及推动包容性变革中能够贡献不可替代的力量。 ## Notable quotes > [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) it's just a war right out there — Margo Georgiadis > [01:15](https://www.youtube.com/watch?v=wDLGYWBalvs&t=75s) make and ship things quickly — Margo Georgiadis > [04:11](https://www.youtube.com/watch?v=wDLGYWBalvs&t=251s) you're going to have to be on a beeper 24 hours a day, seven days a week — Margo Georgiadis > [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) 你们真的愿意直接放弃劳动力市场中 30% 到 40% 最顶尖、最专注的人才吗？ — Margo Georgiadis > [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) put yourself in the path of opportunity — Margo Georgiadis > [11:57](https://www.youtube.com/watch?v=wDLGYWBalvs&t=717s) 你认为我合格吗？如果不合格，我需要提升哪些方面？ — Margo Georgiadis > [14:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=870s) 我很高兴你提到随时待命。事实上，我今天的工作产出已经超过了大多数男性同行。如果真正有必要，我非常乐意待命；但比起形式，我更看重的是实际影响力与工作价值。 — Margo Georgiadis > [14:30](https://www.youtube.com/watch?v=wDLGYWBalvs&t=870s) put the miles on the feet — Margo Georgiadis ## Mentioned Emma Hinchliffe, Margo Georgiadis, Fortune, Montai Therapeutics, Ancestry, McDonald's, Mattel, Google, AppLovin, xAI (Grok), Tesla, Mira Murati, Thinking Machines Lab, OpenAI, DeepSeek, Fidji Simo, Instacart, ChatGPT (GPT-5), Gemini, Ai4 ## Video description Emma Hinchliffe, Editor, Fortune Most Powerful Women, Fortune Margo Georgiadis, Co-Founder & CEO, Montai Therapeutics Abstracts A candid fireside chat with leading women in AI, exploring their journeys, insights, and impact on the future of the field. [Source](https://www.youtube.com/watch?v=wDLGYWBalvs)

Source quotes

[1:15] Margo Georgiadis: it's just a war right out there [1:15] Margo Georgiadis: make and ship things quickly [4:11] Margo Georgiadis: you're going to have to be on a beeper 24 hours a day, seven days a week [11:57] Margo Georgiadis: 你们真的愿意直接放弃劳动力市场中 30% 到 40% 最顶尖、最专注的人才吗？ [11:57] Margo Georgiadis: put yourself in the path of opportunity [11:57] Margo Georgiadis: 你认为我合格吗？如果不合格，我需要提升哪些方面？ [14:30] Margo Georgiadis: 我很高兴你提到随时待命。事实上，我今天的工作产出已经超过了大多数男性同行。如果真正有必要，我非常乐意待命；但比起形式，我更看重的是实际影响力与工作价值。 [14:30] Margo Georgiadis: put the miles on the feet

Source quotes

[1:15] Margo Georgiadis: just a war [1:15] Margo Georgiadis: probably the biggest $150 billion company you've never heard of [4:11] Hiring Manager: And you're going to have to be on a beeper 24 hours a day, seven days a week [4:11] Margo Georgiadis: bro culture sexism [4:11] Margo Georgiadis: slinging beers at the bar afterwards [11:57] Margo Georgiadis: Remind yourself that if you don't put yourself in the path of opportunity and let people know what you're interested in and ask them, 'Do you think I'm qualified? And if I'm not, what do you think I need to do?' If you don't make that effort, you don't remind them that you want that. [14:30] Margo Georgiadis: Well, I'm glad you talked about wearing the beeper. I probably outwork most of my male colleagues today. Happy to do a beeper if it's really necessary. What I really care about is impact and doing the work that I want to do. [14:30] Margo Georgiadis: the miles on the feet, on the phone, on the video conferences

Original transcript

[0:04] Hi Marggo. How are you? [0:05] >> I'm great Emma. It's so great to see all [0:08] these awesome women. [0:09] >> I know. It is so exciting to see people [0:11] filling up this room and all you guys [0:12] coming in. I mean, walking around this [0:14] conference, it's like we're all here. [0:15] So, thank you for coming. [0:17] So, Margot, you are the former CEO of [0:20] Ancestry. You've worked across [0:22] organizations and on boards from [0:24] McDonald's to Mattel to Google. Now, [0:26] you're building Montai. So just really [0:28] quick tell us about what Montai is and [0:29] what you're building. [0:30] >> Sure. Um we're obsessed about creating [0:34] uh small molecule therapeutics for [0:36] chronic disease. So we saw this huge [0:39] problem that we have all these really [0:41] innovative therapies in life sciences [0:43] but they don't actually get to broad [0:45] populations. So even 30 50% of the most [0:49] severely diagnosed patients don't have [0:52] access to life transforming medicines. [0:53] less than 10% of all diagnosed patients. [0:57] So we created an idea leveraging AI and [1:00] reimagining how we discover drugs to be [1:03] able to solve that problem. So really [1:05] excited to take our first four therapies [1:07] towards the clinic and I don't come from [1:10] life sciences so trust me that was a [1:12] tough industry to learn. [1:15] That's amazing. Um so you have a view [1:17] into like these AI native companies as [1:19] well as legacy businesses that are [1:20] figuring out how to integrate AI right [1:22] now. So if you could kind of break down [1:25] the unique challenges faced by both [1:26] kinds of companies, let's maybe start [1:28] with AI native startups. What are the [1:30] biggest challenges for those businesses [1:32] right now? [1:33] >> Yeah, I mean I think if I think about AI [1:36] natives, so I mean foundation model [1:37] companies or um AI focused software [1:40] applications. I think we all see uh that [1:44] it's just a war right out there. Um and [1:47] the question is how many can actually [1:49] survive um in those and I think more and [1:51] more people are starting to shift to the [1:53] applications layer and trying to think [1:56] about which of those will be resilient [1:58] businesses or will the foundation models [2:01] eventually right transfer into a lot of [2:04] those areas and I think a lot of that's [2:06] going to be determined on distribution [2:09] um if you get a toll hold in many [2:12] industries and you are the application [2:14] of record in many indust ries, it's [2:17] really hard to take you out. And I I [2:19] think that's particularly true in areas [2:21] like um healthcare, you know, where [2:24] people are really really sensitive to [2:26] small differences and so they get used [2:28] to it. Um but at the same time, I think [2:30] we have a big issue in terms of the war [2:32] for talent. um you we all see these [2:35] crazy pay packages going around and one [2:38] of the things you know being involved in [2:39] several of these companies as an adviser [2:42] or early independent board member that [2:44] I'm excited about is I feel like [2:46] companies are realizing that it's not [2:49] just about the pay it's also about the [2:51] culture we create. So, for example, at [2:53] App Leven, probably the biggest $150 [2:55] billion company you've never heard of. [2:57] Um, it's a challenger that's reimagined [2:59] how to do ad serving, but I really [3:01] admire the engineering team because what [3:04] they really said is we want to create an [3:06] environment where young people want to [3:08] come here the best and the brightest [3:09] because they get to make and ship things [3:11] quickly and we trust them and we [3:13] engineer an environment that makes them [3:16] feel valued. And so that enables them to [3:19] disproportionately attract some of the [3:21] top talent away from Grock, Tesla, all [3:24] these other companies because they want [3:26] to work somewhere where when they show [3:28] up at work, somebody cares that they're [3:30] there. [3:32] They actually let them even interns. I [3:34] was at the intern fair last week after a [3:37] board meeting and these people were so [3:39] excited because as interns, they [3:41] actually shipped things. They didn't [3:42] just get people coffee or hang out. they [3:45] actually worked on real meaningful [3:46] things and saw their impact. And so I [3:48] think there's a lot we need to do as [3:49] organizations to do that. And as we [3:51] think about women and recruiting women, [3:53] women disproportionately care about [3:55] impact and working somewhere they feel [3:57] like they belong more than just getting [4:00] a big pay package. And I think if we're [4:02] going to bring more women talent into uh [4:05] the those core AI companies, I think we [4:07] have to think differently about how we [4:08] recruit, attract, and promote talent. [4:11] >> That's such an important point. And I [4:13] think, you know, we just saw an [4:14] interesting case study of this when you [4:16] look at M. Marotti and thinking machines [4:18] lab and the, you know, billion-dollar [4:21] offers that went out to some of her team [4:22] members who then chose not to take them [4:24] and that, you know, I think, do you [4:26] think that culture played a big part in [4:27] that? What did that tell us about how [4:29] women are building in AI and what women [4:32] might value in AI? [4:34] >> Yeah, it's really interesting. You know, [4:36] I I coach a lot of women technical [4:39] leaders coming up the ranks and I'm [4:41] constantly shocked at some of the [4:44] experiences they have interviewing. What [4:46] a recent story, which you guys can all [4:47] raise your hands if you've ever had this [4:49] happen to you. Um, super top talent [4:52] product and engineering lead goes to [4:54] interview at one of those big foundation [4:55] companies that was recruiting her. But a [4:58] big piece of the interview was the guy [5:00] essentially saying, "And you're going to [5:01] have to be on a beeper 24 hours a day, [5:04] seven days a week." And that was his way [5:06] of saying, "I want you to be committed." [5:08] But that's not really resonating with [5:10] her at all. Like she just stopped the [5:12] interview there and walked away. And he [5:14] was just completely tonedeaf that she [5:17] wants to have impact. If he would have [5:18] talked her about impact, she probably [5:20] works harder than any of the male guys [5:21] she works with today. But the way he [5:24] approached her was so off-putting that [5:26] he lost that talent completely [5:28] fundamentally because what that said is [5:30] what we see 70% of women seeing is they [5:33] feel that bro culture sexism. They don't [5:36] feel like, you know, if they can't, you [5:38] know, put up the hours or be around with [5:39] the tough guys, you know, slinging beers [5:41] at the bar afterwards, they're not going [5:43] to be accepted. And that's not the kind [5:45] of culture that's going to attract, [5:46] develop, and retain, right? That huge [5:48] portion of talent. And these companies [5:50] can't be successful without being able [5:52] to recruit a broader, you know, set of [5:54] candidates. So, I think there's a lot we [5:56] can do as leaders. [5:58] It's how do we write job descriptions so [6:00] they're more inclusive? How do we [6:02] actually use AI as part of the [6:03] opportunity so that we make sure that [6:05] we're not just recruiting people in [6:07] their own image, but we're actually [6:09] pushing the thinking about what are the [6:11] skills that are needed to be in the room [6:13] uh so that we can be successful. And [6:15] then we should look at candidate success [6:16] rates through the pipeline and who [6:18] interviewed them. We learned a lot when [6:20] I was at Google doing that analysis. We [6:23] realized certain people that just [6:24] shouldn't be on interview slates because [6:26] every diverse candidate person of color [6:29] never made it past them. They always got [6:31] low scores. And so you start being able [6:33] to detect some of those cycles. And [6:34] that's where AI can be actually used for [6:36] good. [6:38] >> So you're hiring, you're trying to bring [6:40] more women into your workforce. You say [6:43] women value impact. you know, if you get [6:45] kind of really granular about it, what [6:46] does that look like? What is the pitch [6:48] that you're making to them? I mean, in [6:49] the role you're in now, I think the [6:51] impact is clear. You're in biotech. It's [6:53] obviously impacting humanity, human [6:55] health, but let's say you're in a [6:57] company that doesn't have quite that [6:59] obvious of an impact element. How would [7:00] you make that pitch? [7:02] >> Yeah, I think I I would really just [7:04] focus and I that's why I gave the [7:05] example of Apploving, right? Because [7:08] they're doing ad serving, but it's a [7:12] giant AI engine. and they're like a [7:13] deepseek. And in that case, that was [7:16] really about you get to come and work [7:18] somewhere where your talent is valued, [7:20] your creativity is valued, and we're [7:22] encouraging you you to be in a flat non- [7:24] hierarchical organization [7:27] where you know the senior leaders, they [7:29] actually care about you, and they're [7:31] motivating you to take risks to innovate [7:34] at this company. And I think that's [7:35] super appealing also to actually care [7:39] about people. I was amazed at the head [7:41] of engineering for some of his teams in [7:43] a foreign country, he actually gave a [7:46] 4-day holiday and you took your parent. [7:48] You had to take your parents because in [7:50] that culture it was like really [7:51] important to them. So just showing that [7:54] you actually care about people at a [7:55] human level goes a really long way. [7:59] >> So in general obviously as we know women [8:02] are underrepresented in AI but to you [8:04] why does that matter? Why is that [8:05] important? Yeah, I think a lot of people [8:08] I've been in a lot of dinners where [8:09] people have this debate, you know, are [8:11] we just in a war where we just have to [8:12] do whatever we can to stay ahead in [8:15] these industries and does this not [8:16] matter? I would say, you know, we're [8:18] sitting at this conference at AI4 in an [8:21] inflection point in this industry where [8:24] this technology is capable of [8:26] reimagining our work, industries, [8:29] people's lives in such a fundamental [8:30] way. how we design and architect those [8:33] products will have such a fundamental [8:36] impact on people's lives that we have to [8:38] ensure that the way we do that is [8:39] inclusive. So you gave the example of [8:41] Mera which I think is great in terms of [8:43] open source. I'm also excited to see [8:45] Fiji going Simo going to open AI and [8:49] actually creating architecture that's [8:52] looking at not just technical outcomes [8:54] but also human outcomes and how those [8:56] two things are connected because at the [8:58] end of the day how people see the value [9:02] of these products and experience them at [9:04] a personal level will have a big impact [9:06] on their adoption and we've all those of [9:08] us who've worked in AI how many people [9:10] in this room have worked in AI for like [9:12] 10 20 years my passes a lot of people [9:14] here. We've seen these things go wrong [9:17] even in the last four or five years, [9:19] right? Think about predictions and [9:20] policing. Think about health care where [9:23] we know that in life sciences less than [9:25] 1% of the research was actually done on [9:28] women's specific diseases. We know that [9:30] diverse groups that don't get regular [9:32] access to health care when you look at [9:34] clinical data. There's fundamental bias [9:36] in the data that you're looking at. We [9:38] see this in recruiting where the bots [9:41] will choose certain types of résumés. If [9:43] you base them on the resumes you already [9:45] have, it's going to go and select people [9:47] that look exactly like that, not [9:49] necessarily consider a broader range of [9:51] skills. I could go on and on. We know [9:53] these things are real and AI can either [9:56] be a challenge or we can turn it around [9:58] and say where are those risks and how [10:00] are we going to use it to our advantage? [10:01] And that's where I see the really great [10:03] companies embracing this in a way that [10:05] it's an imperative, right? They're [10:07] setting OKRs for this because it's a [10:10] business objective. They want their [10:11] products to be the best they can [10:12] possibly be and they know that if they [10:14] don't have diverse people included in [10:16] those conversations that that's not [10:18] going to happen. [10:19] >> How can companies measure that? [10:22] >> So, I think that it's really important. [10:24] You know, I used to do this. I think a [10:25] really good example that everyone could [10:27] relate to is when I was at Ancestry, we [10:29] would do a lot of innovation with AI and [10:31] both new health offerings but also [10:33] family um history experiences. Well, [10:36] depending what your family history's [10:38] experience was coming to this country, [10:41] how you might experience some of these [10:43] products would be very different based [10:45] on your ethnic background. And it was [10:48] extremely important every time we [10:50] created something new that we had a [10:52] diverse set of people that would [10:53] actually think about how their product [10:55] would work and how the UI would be [10:57] experienced by those individuals so that [10:59] we were being respectful also how we [11:02] marketed them and took feedback on them. [11:05] That's really possible to do in any of [11:07] the categories that these companies [11:09] exist in. And I it's just as true for a [11:12] consumerf facing service as it is a B2B [11:15] because when B2B services are actually [11:18] having people feel like their work could [11:20] actually be eliminated, there can be [11:22] enormous cultural resistance and [11:23] barriers. I saw this a lot in life [11:25] sciences where people were trained, you [11:28] know, 20 years sometimes on a cell type [11:29] in a very narrow skill area. So where [11:32] you have high skill specialization, [11:34] people tend to be very resistant on [11:36] adopting new tools that are very [11:38] multiffactorial because for them they [11:40] feel like wow I only know this one thing [11:42] so what am I going to do in a world [11:45] where all these things are integrated [11:46] together and how do we help people [11:48] understand that so I don't think these [11:50] integrative challenge or these [11:52] inclusivity challenges are any different [11:54] in B2B versus B TOC. [11:57] So the point of view you mentioned [11:59] people who are probably you know in [12:01] other rooms of this conference right now [12:02] who feel that you know this is a war [12:04] moment in AI like it's all hands- on [12:06] deck we don't have time to consider this [12:07] type of thing like how do you convince [12:10] them otherwise? I mean you've shared [12:11] some of your arguments but let's say [12:12] you're one-on-one with a person who [12:14] really feels that way. You're [12:15] interviewing with a company that feels [12:16] that way. You're working at a company [12:17] that's operating that way. What can you [12:19] do? [12:20] >> Yeah. I mean I'm in those rooms and [12:22] that's when I remind them what an [12:24] important share of the workforce is. So, [12:26] are you really willing to walk away from [12:28] 30 40% of the most talented people and [12:31] in my experience some of the people that [12:32] will be the most dedicated [12:35] um to the future? And then they'll ask [12:37] me, well, how do I attract them? How do [12:39] I think differently about this? And then [12:41] I give them a lot of those examples. And [12:42] I think a lot of it they get stuck, [12:45] right? Because if you are in that [12:47] organization that's all guys at the top [12:49] who are just thinking about what they're [12:51] comfortable with, they don't realize you [12:53] don't need to contort yourself. But you [12:55] do need to do some things differently. [12:56] And sometimes I do have to have hard [12:58] conversations. Like if there are no [13:00] women at the top, we see this [13:01] systematically in every company I've [13:03] ever worked in. If there's no woman [13:05] ahead, women often will opt out of that [13:08] company because they don't see a path [13:10] that's reasonable. So you do have to go [13:13] the extra mile, you know, and and [13:16] sometimes I'm like, hey, if you really [13:18] care about that portion of the [13:20] population, it's worthy of your time. It [13:22] doesn't take a lot of time. You just [13:24] need to let that person know that you [13:26] value that they're there and that you [13:28] help think about their career because [13:30] women on balance in my experience like [13:32] the guys are all running to my office [13:34] telling me that they are really [13:36] qualified for the next opportunity or [13:38] what that new thing is they want to [13:39] learn. The women on balance do that way [13:41] less. And so if you're not consciously [13:44] realizing that's true, you will every [13:46] time you think about promotions or the [13:48] cool assignment, you will always select [13:50] the guys because they're the ones that [13:52] were coming to your office. Now women, [13:54] all of you in this room, remind yourself [13:56] that if you don't put yourself in the [13:57] path of opportunity and let people know [14:00] what you're interested in and ask them, [14:01] do you think I'm qualified? And if I'm [14:03] not, what do you think I need to do? If [14:05] you don't make that effort, you don't [14:07] remind them that you you want that. And [14:10] unfortunately, the guys are doing that [14:11] all the time. And so you do need to we [14:13] need to both take more action as women [14:15] but I think men need to be sensitive to [14:17] the fact that people from different [14:19] cultures right and women tend to be [14:21] biased not to ask for things and so [14:24] there's very small changes that I think [14:26] people can make in leadership that can [14:28] make a huge difference. [14:30] What advice do you have for women who [14:32] are maybe not yet in leadership at a [14:33] more junior level who are maybe like [14:35] that person you described who are on the [14:37] interview circuit and having experiences [14:38] that are making them question is this [14:40] the right path for me? is this am I [14:42] going to find the right fit for myself [14:44] in this industry even if they're very [14:46] enthusiastic about AI how can they [14:48] figure out what the right fit is for [14:49] them and make sure that their talents do [14:51] not go to waste in this burgeoning [14:53] industry I do think it's tricky right [14:55] even that executive that I was telling [14:57] you earlier about that I was mentoring [14:59] who had that experience about the beeper [15:01] you know one of the things that I you [15:03] know suggested that she say next time [15:06] that happens to her is just to turn it [15:08] around on him like don't take it [15:10] personally but lean back in and say, [15:13] "Well, I'm glad you talked about wearing [15:14] the beeper. I probably outwork most of [15:16] my male colleagues today." Um, happy to [15:18] do a beeper if it's really necessary. [15:20] What I really care about is impact and [15:23] doing the work that I want to do. And [15:24] send a message back that I hear you, but [15:27] you're also talking about a culture or [15:30] really ask, you know, why is that? Are [15:32] you testing me because you want to know [15:34] if I'm committed? If it's that, I'm in. [15:37] But if you're just saying I have to wear [15:38] a beeper just because I have to, maybe [15:40] that isn't a place, right, that you [15:42] really want to work. And I I do think [15:43] women do want to choose environments [15:46] where they are going to feel personally [15:48] comfortable. And that won't be [15:49] everywhere. Um, and I'm just being [15:51] honest about that. Um, I would self- [15:53] select into those places that do share [15:56] your values, but I think the best way to [15:58] find those is through personal networks. [16:02] men and women that you know really [16:04] understand what that makes that [16:06] organization tick. How do they think [16:08] about talent? How do they develop [16:09] talent? And not every organization is [16:11] right for everyone. Um but I think you [16:14] can find that fit but you have to put [16:15] the work in. You got to put the miles on [16:17] the feet on the phone on the video [16:19] conferences with everyone that you know. [16:22] >> So let's talk for a moment about the AI [16:25] integrating businesses legacy [16:27] businesses. What are you hearing from [16:29] women in executive roles throughout [16:31] those industries? Are you seeing a [16:32] confidence gap in AI adoption and how [16:35] can we begin to close it? [16:37] >> Yeah, there is unfortunately is a big [16:39] confidence gap. I mean I the recent most [16:41] recent survey I saw, thank you chat GBT. [16:44] I was looking up before I came here. G T [16:46] G T G T G T G T G T G T G T G T G T GB5 [16:47] is really good. I was really impressed [16:49] and I stress tested it against Gemini [16:51] after that. Um, and they were both good, [16:53] but I would give the edge to chat five [16:55] on this one, even though I'm an [16:56] ex-gooller. Um, but 80% of the women are [17:01] in professional roles are saying they [17:03] feel like they don't get enough [17:04] education in AI. And I think if we think [17:07] about women's confidence gaps in [17:09] general, I think in the broad [17:10] population, right, 30% of women roughly [17:14] are comfortable with AI and what it [17:15] means and 71% of men. So we know that [17:18] there's a divide and we need to find a [17:20] way to close that gap because I think [17:22] that will limit the talent um that gets [17:24] into the arena. But in those industries [17:27] what I what I really believe and I've [17:28] been part of is at the sea level I think [17:31] leaders need to commit to really getting [17:34] educated about AI. I think they have to [17:36] spend between two and three hours every [17:39] week given the pace of change and the [17:41] size of the opportunity [17:43] either reading in areas that are [17:46] relevant to your business, talking to [17:48] other people in the industry that are [17:51] working in similar areas or just working [17:53] on the tools yourself. Actually use them [17:55] and stress test them. Ask them the [17:57] hardest question because unless you [17:59] immerse yourself, I think it's really [18:00] difficult to lead this change and when [18:02] it's not led by the seauite, I think [18:03] it's not effective. I think second um in [18:07] those companies I think we really have [18:08] to think about those executives that are [18:10] really good change integrators and often [18:14] that tends to be women. These are people [18:15] that are deep domain expertise on the [18:17] way things are, but they also can see [18:19] possibilities and they have the empathy [18:22] around how do we think about change and [18:25] what's possible in an organization and [18:27] how do we help people digest that and [18:29] reset the way we work and the way we [18:31] measure success so that people feel like [18:33] they're included because when they're [18:35] not, it's very hard to overcome that [18:37] cultural resistance. So, I see a lot of [18:39] work going on in those areas and I I [18:41] think we as women can be a big part of [18:43] that change. [18:44] >> Well, thank you. Thank you so much, [18:45] Margot, for an amazing conversation and [18:46] thank you all for being here. [18:48] >> Thank you.