Colgate, SOCi CMOs Pivot to Brand Moats as AI Automates Campaign Work

AI实现营销自动化，Colgate与SOCi CMO转向构筑品牌护城河 随着生成式AI将常规内容创作商品化并压缩产品上市周期，营销高管警告称，技术速度已成为基本门槛，独特的人性化品牌标识以及人均ARR已成为唯一定位稳固的竞争护城河。 营销领袖们分享了在大型企业及成长期组织中推进AI落地的实用策略，详细阐述了构建统一内部“公司大脑”以及将产品上市周期从数月压缩至数周的具体实践。专家组探讨了搜索与发现机制的结构性演变，包括从传统SEO向答案引擎优化（AEO）的转型，以及在实现自动化之前审计工作流的必要性。他们强调，尽管AI有效消除了重复性的行政事务，但真正的竞争优势依然依赖于保留真实的人文叙事与品牌灵魂。最后，讨论概述了现代衡量框架，敦促营销团队摒弃追溯性的营销组合模型，转而通过人均ARR等董事会级别的效率指标来传递价值。 当AI立足于清晰规范的数据分类与集中的机构内部“隐性知识”，而非孤立的执行工具时，方能发挥其最大的战略优势。 (5:48) 将分散的企业数据汇聚到集成于Slack的统一“公司大脑”中，可使团队即时提取客户沟通洞察，并实现递归式自我修正。 (11:14) 将上市周期从数月压缩至数周需要大量前期工作流工程与提示词微调，而非单纯依赖初级的单次提示词AI输出。 (14:40) 从传统SEO向答案引擎优化（AEO）的转变，要求跨去中心化数字资产规范结构化Schema与数据可信度。 (16:10) 组织在接入AI能力之前必须先优化内部工作流与部门架构，以避免单纯加速原本就存在缺陷的流程。 (17:56)

Enterprise marketing leaders argue that while agentic AI drastically compresses operational cycles and democratizes organizational knowledge, sustainable advantage relies on rigorous workflow restructuring and distinct brand positioning. As technological speed becomes commoditized table stakes and boards benchmark efficiency by revenue per employee, authentic human storytelling and brand governance serve as the primary defensible moats. Data Architecture and Centralized Enterprise Knowledge (11:14) Panelists agree that effective AI deployment requires clean data taxonomy and integrated endpoints. Kathleen Booth detailed Sequel.io’s internal system, 'Cerebro,' which connects CRM data, product telemetry, call recordings, Intercom chats, and documentation into a Slack-based interface. Using Supabase and vector search for semantic retrieval alongside recursive CTO corrections, the system surfaces hidden tribal knowledge and assists customer support in real time. Monica Ho of SOCi cautioned, however, that integrating sparse or incomplete external datasets into internal systems creates acute risks of AI hallucinations. At Goodwill Industries International, Onney Crawley described an enterprise audit across 150 autonomous local organizations to standardize schema, backlinks, and citations. This unified data baseline addresses the industry shift from traditional search engine optimization (SEO) to answer engine optimization (AEO), ensuring generative AI search engines cite credible corporate facts rather than bypassing contradictory or unverified local websites. Compressing Operational Cycles and Workflow Hygiene (14:40) As software engineering velocity accelerated with AI, marketing emerged as the organizational bottleneck. Monica Ho noted that SOCi compressed its go-to-market enablement and collateral production cycles from 8–12 weeks down to 1–2 weeks by deploying specialized AI modules. However, Ho stressed that production-grade outputs require months of prompt engineering, workflow structuring, and human validation rather than naive, raw prompting. Technology integration must follow structural refinement rather than precede it. Onney Crawley emphasized that Goodwill audited departmental priorities and eliminated friction prior to tool adoption, warning that automating disordered processes merely accelerates inefficiency. Similarly, Christy Borrowman highlighted Colgate-Palmolive's focus on eliminating low-value mechanical repetition—such as marketers manually copying identical product claims across 35 disparate systems including digital asset management (DAM) and product information management (PIM) platforms. Talent Reallocation and AI-Native Hiring Standards (24:04) Marketing roles are shifting from manual content execution to AI agent architecture, prompt fine-tuning, and output supervision. Monica Ho noted that SOCi evaluates candidates on their direct experimentation with frontier models like ChatGPT and Claude to build automated workflows. Concurrently, Onney Crawley identified a generational split among junior applicants: traditional graduates who viewed AI assistance as cheating versus AI-native interns who rapidly outperformed senior peers on tactical output. To compress onboarding timelines for revenue-generating roles, Kathleen Booth outlined Sequel.io's partnership with SV Academy to hire pre-trained Business Development Representatives (BDRs). Candidates complete cold-calling and objection-handling bootcamps and undergo a three-month trial embedded in the marketing team without upfront recruiter fees, allowing new hires to book sales demonstrations within their first week. Measuring Performance Through Real-Time Efficiency and ARR (32:38) Traditional retrospective analytics, such as quarterly Marketing Mix Modeling (MMM) or click-through rates, are insufficient for automated campaigns operating at high velocity. Christy Borrowman urged a shift toward real-time predictive feedback loops, cautioning against anecdotal executive evaluations based on informal ChatGPT queries. In venture-backed environments, Kathleen Booth differentiated between internal leading indicators and board-level lagging indicators. While internal teams benefit from operational efficiencies—such as Sequel.io reducing webinar production time by roughly 80%—executive boards prioritize macro-level capital discipline. Facing valuation benchmarks set by lean AI startups like Cursor and Lovable, boards increasingly evaluate marketing performance through Annual Recurring Revenue (ARR) per employee and revenue productivity per sales representative. Brand Soul as the Core Defensible Moat (33:45) As access to generative tools universalizes content volume and speed, panelists maintain that brand identity becomes the ultimate commercial differentiator. Christy Borrowman cited Glean's principle—'Don't automate the soul out of work'—advocating that automating administrative busywork frees human practitioners to focus on emotional resonance and product differentiation. Onney Crawley and Kathleen Booth argued that marketing is experiencing a return to strategic brandcraft reminiscent of the 'Mad Men era.' To combat an influx of generic AI-generated content across digital platforms, organizations must ground automated external messaging in a centralized, human-directed 'brand brain' that aligns customer feedback with the company's core strategic narrative. The panel concludes that as AI transitions from a novel differentiator into standard operational utility, marketing success will depend on disciplined data foundations and distinct brand positioning rather than tool adoption alone. However, enterprise leaders continue to grapple with the manual burden of data preparation pipelines, persistent cross-platform data conflicts, and the challenge of proving that operational time savings directly expand top-line revenue.

多位企业营销高管指出，人工智能虽能通过打通内部数据孤岛、大幅压缩上市周期并剥离低价值机械工作来提升运营效率，但随着技术工具的快速普及，真正的长期竞争壁垒正从单纯的内容生产速度转向无法被算法复制的品牌定位与真实叙事能力。 统一数据底座与企业大脑激活隐性知识 (11:14) 营销效能的释放高度依赖底层数据的规范与打通。Colgate-Palmolive的Christy Borrowman指出，缺乏系统化的数据分类体系与内部接口，AI扩增便无法发挥实质作用。Sequel.io的Kathleen Booth分享了其内部构建的“Cerebro”系统：该系统通过打通CRM、营销自动化、销售与客服通话记录、Intercom会话以及内部HR维基等多方数据，在Slack内建立起统一的企业大脑。该工具不仅能提炼真实的客户评价用于信息定位，还能在回答中呈现置信度评分；当技术负责人或高管在Slack讨论串中修正错误时，系统会执行递归学习以持续优化输出精度。Booth进一步说明，团队利用Supabase结合向量搜索，能够按照语义语境检索非结构化客户反馈，克服了传统精确关键词查询的局限。 SOCi的Monica Ho则从外部信号补充说明，除企业第一方数据外，聚合本地搜索趋势与顾客评价等微观信号至关重要。但她同时警示，盲目接入未清洗或稀疏的数据源会显著加剧AI生成幻觉的风险。 交付周期压缩与从SEO向AEO转型 (14:40) 随着研发端借助AI加速软件交付，营销团队往往成为整体业务交付的新瓶颈。SOCi的Monica Ho表示，传统新产品上市（GTM）往往需要在产品营销与物料制作上耗费8至12周时间；通过数月对提示词的精调、模块化工作流搭建以及人工监督校准，SOCi成功将上市落地周期压缩至1至2周。Ho强调，可靠的企业级AI应用需要前期大量的流程重构，简单的基础提示词无法直接产出符合业务要求的物料。 在外部传播渠道上，搜索逻辑正从传统的搜索引擎优化（SEO）快速迈向答案引擎优化（AEO）。Goodwill Industries International的Onney Crawley指出，AI引擎更倾向于直接生成综合摘要而非呈现链接列表。Goodwill旗下150个独立运作的分支机构各自运营网站，分散且矛盾的信息容易被AI引擎视为不可信来源。为此，Goodwill对其庞大的网络资产开展全面抓取与审计，统一结构化模式与引用规范，以建立权威的数据基线，确保AI答案引擎生成准确的企业事实并消除外界误解。 组织流程重构与营销人才技能分化 (21:55) 在部署AI工具之前，企业必须优先理顺组织内部的工作流。Goodwill的Onney Crawley强调，盲目引入AI只会加速原本低效或错误的运作。在团队重构中，Crawley观察到年轻应聘者呈现两极分化：部分应届毕业生视AI工具为作弊手段而产生抵触，另一些“AI原生代”则凭借熟练的提示词与工具运用在产出效率上超越资深员工。若资深人员长期沿用固有模式，可能会延缓团队的智能化转型。 Colgate-Palmolive的Christy Borrowman指出，企业营销人员过去耗费大量时间在数字资产管理系统（DAM）与产品信息管理系统（PIM）等多达35个分散系统中重复复制粘贴同一文案，AI的核心价值在于剥离此类低价值行政琐事。Sequel.io的Kathleen Booth补充道，其营销部门承担100%的销售线索指标，通过与SV Academy合作引入经过外呼与异议处理强化的业务拓展代表（BDR），实现了新员工入职首周即敲定演示的高效落地。SOCi的Monica Ho表示，营销岗位正从单一渠道的机械执行，转变为流程架构设计、提示词调优以及幻觉审查等高阶职能。 衡量指标演进：从滞后归因到人均ARR (29:01) 传统的营销效果评估方式正在经历重塑。Colgate-Palmolive的Christy Borrowman指出，以往依赖按季度或年度回溯的营销组合模型（MMM）以及点击率等滞后指标，无法支撑当前AI驱动的高频内容分发与实时个性化互动，营销部门需要建立面向未来的实时预测性反馈机制。她同时指出，管理层不应依赖在ChatGPT中随意检索“最佳牙膏”这类偶发且缺乏统计意义的虚荣查询来评判营销成果。 在资本环境收紧的背景下，Sequel.io的Kathleen Booth将AI成效区分为领先指标与滞后指标：内部活动制作耗时缩减80%属于效率维度的领先指标，而董事会真正聚焦的是人均年度经常性收入（ARR per employee）等营收结果。由于投资者常以Cursor或Lovable等高人效AI原生企业作为参照标杆，营销负责人需向管理层清晰证明，团队能够在不按比例扩充编制和预算的前提下驱动业务规模化增长。 工具套利均质化下的品牌灵魂护城河 (33:45) 随着生成式AI逐步演进为底层基础设施，技术本身带来的效率优势正在迅速转变为行业标配。Colgate-Palmolive的Christy Borrowman援引观点强调，企业在推进自动化时不应抹杀工作中的“灵魂”，营销人员省下低效精力后，应聚焦于构建具有情感共鸣与科学产品优势的差异化品牌认知。 Goodwill的Onney Crawley认为，AI终将像电力或互联网一样融入日常运作，而不再作为独立概念被频繁提及。Sequel.io的Kathleen Booth总结指出，营销在本质上是一场寻找竞争优势的套利游戏；过去十年基于自动化速度的技术套利空间正随着工具普惠而收窄，真实、清晰且具有辨识度的品牌叙事将成为AI时代无法被算法替代的核心护城河。 本场研讨展现了跨国快消、非营利组织及初创SaaS企业将AI融入数据基建与工作流的具体路径，但小组未进一步探讨当全行业普遍依赖AEO与算法生成内容时，企业在长期面对内容同质化时的应对机制，以及非结构化企业大脑在复杂合规场景下的治理边界。

## Background 本内容由 Ai4 于 2026 年 8 月 22 日发布在 YouTube 平台上。视频记录了一场题为“精明扩张：AI 如何驱动营销效率与精准度”的圆桌研讨会，汇集了来自 ISS、Goodwill、高露洁-棕榈、Sequel.io 及 SOCi 的营销高管，共同探讨 AI 在创意自动化、上市周期压缩、数据治理及品牌真实性等维度的重构与实践。 ## TL;DR 本场研讨会深入剖析了 AI 时代营销团队如何实现从执行者向系统架构师的范式转型。嘉宾们分享了通过构建“企业大脑”整合全渠道数据、将 GTM 周期从数月压缩至两周、以及应对零点击搜索（AEO）的实战案例。讨论还指出，随着技术红利逐渐成为行业标配，企业需从滞后归因转向即时预测，并最终回归以人为本、情感共鸣驱动的经典“品牌套利”护城河。 ## Atmosphere & style 整场研讨会节奏紧凑、干货密集，洋溢着务实与清醒兼备的行业探索氛围。主持人的提问直击业务痛点，台上的资深 CMO 与高管们既毫无保留地分享了自建智能体与工作流的前沿战果，又毫不避讳地指出了大模型幻觉、繁琐低效系统以及过度生成内容带来的现实挑战。现场互动真诚而热烈，在严谨的技术拆解与生动的业务案例碰撞中，勾勒出营销人面对 AI 变革时的从容与笃定。 ## Key takeaways - [05:48](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=348s) 数据基建与内容分发是 AI 赋能营销的核心支柱，AI 打破了人力产能天花板，使月产数万条短视频及全渠道智能分发成为可能。 - [11:14](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=674s) 通过构建“企业大脑”（如 Cerebro）打通 CRM 与通话录音等非结构化隐性知识，并通过递归学习实现自我纠错与知识库迭代。 - [09:20](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=560s) 告别静态且昂贵的传统用户画像，利用实时市场信号动态适应消费者，同时严密把控数据输入以防止模型幻觉风险。 - [14:40](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=880s) 重构 GTM 入市流程：经过深度的 AI 工作流微调与模块化重塑，全新产品的上市周期可从 8–12 周压缩至 1–2 周。 - [16:10](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=970s) 从 SEO 向 AEO（答案引擎优化）演进，企业需统一并净化底层 Schema 结构化数据与 Q&A 库，争夺 AI 零点击搜索摘要的权威引用。 - [17:56](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1076s) 营销团队角色发生重构，专业人才从耗时的具体执行者，转型为配置 AI 工作流、训练智能体、制定交付标准与纠偏的复合型专家。 - [21:55](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1315s) AI 旨在将营销人员从跨多系统复制粘贴的繁琐杂务中解放出来，促使团队专注于不可替代的“品牌灵魂”与情感共鸣叙事。 - [24:04](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1444s) 人才招募打破对传统工作年限的依赖，更加注重候选人的 AI 原生思维、实战构建产出以及跨行业快速见效的能力。 - [29:01](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1741s) 营销评估体系从滞后性的事后归因（如传统 MMM 模型）转向近实时监控与前瞻预测（Predictive Optimization）。 - [31:16](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1876s) 向管理层及董事会汇报时应区分过程指标与商业结果，资本市场关注的核心在于“人均 ARR”、管线规模与实际营收增长。 - [33:45](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2025s) 当 AI 效率工具普及成为行业标配（Table Stakes）后技术套利红利将消失，营销将回归经典法则，以“品牌本身”作为核心套利支点。 - [40:24](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2424s) 借助向量检索（Vector Search）捕捉碎片化反馈的深层语境，并依托“品牌大脑”校准发声，确保对外传播与品牌战略定位严格一致。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=4s) 圆桌开场与重磅嘉宾阵容介绍 主持人、ISS 北美区首席营销官 Marjorie McCarthy 热情开场，感谢与会者一早赶来参加本次以“精明扩张（Scaling Smart）”为主题的研讨会。Marjorie 简要介绍了 ISS——一家拥有逾 32 万名员工、业务遍及全球近 60 个国家的工作场所体验与设施管理巨头，并随即将话筒交给台上的四位资深营销负责人进行自我介绍： - **Monica Ho（SOCi 首席营销官）**：Monica 介绍 SOCi 是一家专为拥有成百上千家分支机构或门店的多网点品牌（如 Ford、Ace Hardware、Jersey Mike’s）量身打造的 Agentic AI（智能体）营销平台。其核心是通过经过本地品牌训练的智能体，自动处理本地化评价回复、本地页面优化等繁重工作，以帮助品牌在传统搜索、社媒以及新兴的 AI 搜索中抢占先机。在掌舵营销的八年里，她正带领团队践行智能体 AI 颠覆传统营销流程的承诺。 - **Onney Crawley（Goodwill Industries International 首席营销官）**：Onney 以一句“现场有多少人喜欢淘二手货（thrifting）？”开场，引发台下热烈共鸣。她指出 Goodwill 拥有由 150 个深植本地社区的独立机构组成的联邦制架构。大众熟知其二手门店和捐赠网络，但鲜少了解其本质是将销售收入转化为免费的本地就业培训与职业支持服务。面对各自为政的联邦机构，Onney 的核心职责是在讲好职场赋能故事、吸引捐赠与客流的同时，确保 150 个组织以“同一个品牌”协同发声，而 AI 正在这一进程中起到加速与聚合的关键作用。 - **Christy Borrowman（高露洁-棕榈 Colgate-Palmolive 营销与转型副总裁）**：在高露洁深耕 11 年的 Christy，职业生涯始终围绕数字化、数据与媒体转型展开，并于近期刚履新营销部门旗下的“AI 转型”新职位。她专注于打通全链路端到端流程，探索大型快消品牌未来的营销形态，并结合自己在广告代理商及与开发团队紧密协作的丰富经验推动落地。 - **Kathleen Booth（Sequel.io 营销副总裁）**：Kathleen 介绍 Sequel 最初是业界首个能够直接无缝内嵌至企业官网的研讨会平台，如今正蜕变为一个 Agentic Go-to-Market（智能体市场进入）平台。平台通过采集研讨会全生命周期与官网浏览深度融合的第一方行为数据，在联系人与账户级别进行动态评分，赋能 Datadog、Clay、Morningstar 等客户。作为多所初创和早期成长型企业的营销操盘手，Kathleen 当前的核心目标是打破“团队人头必须与营收成正比增长”的传统定律，在坚持“以人为本（human-first）”和品牌价值的前提下，以 AI 原生（AI-native）模式撬动业务规模化扩张。 ```mermaid flowchart TD accTitle: Panelist Backgrounds and Focus Areas subgraph Panelists["圆桌嘉宾与业务核心领域"] direction TB M["Monica Ho / SOCi<br/>多网点 Agentic AI 本地营销"] O["Onney Crawley / Goodwill<br/>联邦制组织品牌统一与公益叙事"] C["Christy Borrowman / 高露洁<br/>快消巨头的端到端 AI 转型"] K["Kathleen Booth / Sequel.io<br/>精简人效与智能 GTM 数据平台"] end ``` --- ### [05:48](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=348s) 宏观视角：当前 AI 为营销带来的最大竞争优势 Marjorie 抛出第一个宏观层面的核心议题：“在当下这一刻，大家认为 AI 为营销带来的最大竞争优势究竟在哪里？”几位嘉宾不约而同地将核心聚焦于“数据”与“内容分发”，但从各自不同的商业维度进行了深刻剖析： Christy Borrowman 首发发言，指出优势体现于两点：其一是**数据基建**。若企业自身的数据底座不够扎实、分类学（Taxonomy）混乱、链路没有打通，AI 便无法发挥增效作用；只有理顺数据结构，AI 才能让更多业务人员轻松调取并利用核心数据。其二是**内容生产的规模与速度**。Christy 半开玩笑地感叹道：“感谢 TikTok，我现在每个月需要生产 10,000 条内容。”面对如此庞大的吞吐量，任何规模的团队单靠人工都不可能实现，AI 在内容扩量上带来了质的飞跃。 Kathleen Booth 对此表示完全共鸣，并补充了两个关键切入点： 1. **沉淀企业隐性知识（Tribal Knowledge）**：以往企业中最宝贵的行业认知与口头沟通（例如销售与客户的通话录音）极易流失。Sequel 最近构建了“企业大脑（Company Brain）”，不仅汇聚 GTM 系统数据，更将通话记录等非结构化口语资产全部向量化，使原本散落的信息首次成为可全员检索利用的竞争壁垒。 2. **重塑内容分发杠杆**：Kathleen 曾在拥有上万名 GTM 高管的全球社群 Pavilion 担任营销负责人，她观察到绝大多数营销团队的通病在于“把 90% 的精力花在内容制作上，却只有极少精力用于分发推广”。AI 首次让营销人员能够低成本将优质内容自动重构、全渠道精准宣发，真正打破了“重生产、轻分发”的恶性循环。 Onney Crawley 进一步从大型组织管理与消费者洞察的角度切入。她指出，传统调研成果往往被束之高阁，而 AI 推动了数据的“平民化”，让去中心化组织里的每个人都能基于同一事实基准迅速做决策。更重要的是**实时动态适应消费者**。以往花大价钱做的受众细分报告往往几个月就脱节过时，而 AI 赋予了品牌与消费者行为同步进化、实时微调精准定向与互动策略的能力。 Monica Ho 则从多门店本地化营销的独特视角发出洞见与警示：企业自有数据固然重要，但视野往往受限。SOCi 的做法是将企业私有数据与实时的本地市场信号（如各地区用户搜索习惯的差异、对具体门店评价的好恶、当地竞对动态）进行深度匹配融合。同时，Monica 提出了一个冷静的行业警示：不要盲目跟风接入各类系统接口或搭建不成熟的应用；如果输入给 AI 的数据不完整、颗粒度不足，极易引发模型“幻觉（hallucination）”，给品牌带来不可控风险。 - **内容吞吐需求爆发**：以 TikTok 等短视频为代表的新媒体矩阵催生了每月数以万计的创意素材需求，AI 成为打破人力产能天花板的唯一路径。 - **打通暗数据与隐性认知**：通过公司大脑（Company Brain）整合 CRM 与通话录音，把原本沉睡的非结构化销售/服务洞察转化为生产力。 - **告别静态用户画像**：用实时动态的数据学习替代昂贵且易过时的年度消费者调研。 - **警惕劣质数据喂养**：调用外部平台或构建智能体时，必须严把数据完整度，防止底层数据污染导致大模型幻觉。 --- ### [10:47](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=647s) 落地实战：从宏观到具体的 AI 营销战术探索 在结束了宏观维度的探讨后，Marjorie 将讨论引向实操层面，希望嘉宾分享各自团队近期发现或正在推进落地的具体 AI 应用战术。Kathleen Booth 率先接管话题，准备深入拆解 Sequel.io 刚刚完成的内部实践…… ### [11:14](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=674s) 打造企业级“大脑” Cerebro：全源数据整合与 Slack 递归学习 Kathleen Booth 兴奋地分享了她在公司最近一次团建黑客松（Hackathon）中的成果。作为一名《X战警》（X-Men）的忠实粉丝，她与团队受漫画中能定位全球变种人的超级计算机“脑波强化机”（Cerebro）启发，为公司打造了一个名为 Cerebro 的“企业超级大脑”。团队梳理并接入了公司内部几乎所有的数据源，包括客户关系管理系统（CRM）、营销自动化工具、产品使用分析平台、客户支持系统、销售与客户成功团队跨两个平台的通话录音、项目管理系统，乃至内部 HR 文档。系统的交互前端直接集成在 Slack 中，员工无论是在公共的 Cerebro 频道还是通过私信均可直接向其提问。 ```mermaid flowchart TD accTitle: Cerebro 企业智能大脑系统架构 subgraph DataSources["多源企业数据接入"] D1["CRM 与营销自动化"] D2["产品使用与客户支持 (Intercom)"] D3["销售与 CS 通话录音 (多平台)"] D4["项目管理与 HR 文档"] end subgraph CoreEngine["Cerebro 核心引擎"] C1["跨数据源上下文关联分析"] C2["递归学习与知识库记忆回写"] end subgraph SlackFrontEnd["Slack 交互与应用场景"] S1["市场营销：客户高频词汇与竞品定位"] S2["实时客服：通话中即时调取精准解答"] S3["管理闭环：CTO 等高管介入纠错调优"] end D1 --> C1 D2 --> C1 D3 --> C1 D4 --> C1 C1 --> S1 C1 --> S2 S3 -- "修正错误回答" --> C2 C2 -- "持续学习进化" --> C1 ``` Cerebro 展现出了打通孤岛数据的上下文整合能力。Kathleen 尝试让其分析“客户在真实对话中如何描述我们的产品”，系统瞬间遍历了 CRM 记录、两套通话录音系统及 Intercom 聊天记录，直接生成了客户提及频率最高的词汇与短语排行榜——对营销人员而言这堪称“文案金矿”。此外，它还能直接产出成败单分析（Win/Loss Analysis）以及针对竞品的差异化优势洞察。更令团队惊喜的是，客户成功（CS）团队甚至开始用它替代传统的知识库检索：在与客户进行支持通话时，客服在 Slack 线程中输入问题，Cerebro 便结合全部帮助文档与历史支持记录秒级给出详尽答案。当遇到不完整或错误的回答时，CTO 等技术专家会直接在线程中指正，系统通过“递归学习”（Recursive Learning）将修正内容回写至记忆库，实现每一次纠错后的自我迭代。 - **接入全量企业数据**：整合 CRM、产品用量数据、客服工单、项目文档、跨平台销售/CS 通话录音及 HR 知识库。 - **营销赋能案例**：自动提炼客户真实用语词频榜、竞品对比分析、成单/丢单核心驱动因素。 - **递归学习机制**：专家在 Slack 线程内实时纠偏，AI 自动更新知识记忆，彻底规避企业大模型常见的事实性幻觉累积风险。 --- ### [14:40](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=880s) 破解 GTM 瓶颈：SOCi 如何将产品上市周期从 12 周压缩至 2 周 Monica Ho 聚焦于科技企业面临的严峻现实：随着研发团队借助 AI 极速提升产品发布与功能迭代的节奏，营销团队反而沦为了业务推进的瓶颈。在传统流程中，针对全新产品的市场推进入市（Go-to-Market, GTM）通常需要 8 到 12 周的时间，涵盖产品营销（PMM）、销售赋能培训（Enablement）以及各类核心内容的撰写与制作。为了破局，SOCi 团队决定将大量赋能与内容生产流程重构为标准化 AI 模块与自动化工作流。 Monica 直言，搭建实用的企业级 AI 工作流绝非简单的“写几句 Prompt 提示词再喂点材料”那么轻巧，团队实际投入了数月时间进行深度微调与工作流搭建。这种前期的大量投入带来了跨越式的效率红利：目前 SOCi 的整体 GTM 周期已经从原本的 8 至 12 周大幅缩短至 1 至 2 周。她特别强调，AI 不是一劳永逸的黑盒，前期必须愿意投入时间调优，并在实际运行中始终保持人类专家的监督与审核。 - **GTM 周期大幅缩短**：全新产品与功能的入市筹备时间由 8–12 周压缩至 1–2 周。 - **工具链重构认知**：打破“简单 Prompt 即可解决复杂业务”的误区，构建模块化 AI 工作流需要数月的精细化工程与微调。 - **人机协同原则**：前端自动化提效的同时，必须保留人类专业视角进行把关（Human Oversight）。 --- ### [16:10](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=970s) 从 SEO 到 AEO：Goodwill 统一 150 个独立站点的品牌“真实数据源” Onney Crawley 介绍了 Goodwill 在数字化与搜索转型中的巨大挑战。Goodwill 旗下拥有 150 个独立运营的成员机构，对应着 150 个独立的官方网站。在搜索技术由传统基于链接排名的搜索引擎优化（SEO）向由 AI 生成摘要驱动的“答案引擎优化”（Answer Engine Optimization, AEO）演进的当下，用户往往不再点击跳转到具体网页，而是直接消费 AI 总结的答案。然而，由于 150 个站点之间存在大量相互矛盾、缺乏佐证且未标准化的分散内容，AI 引擎在抓取时会判定其不可信而直接放弃引用，导致品牌在 AI 搜索环境中严重失声。 为了夺回品牌在 AI 时代的信息主导权，Goodwill 启动了一项庞大的数字工程。团队正在对所有 150 个站点进行全量扫描与爬取，系统性重构网页结构化数据（Schema）、外链结构、问答库（Q&A）以及内容底层设置，旨在将混乱的内容净化为高度一致、经过清洗的官方权威数据源。这不仅能确保用户无论在何地搜索任何一家 Goodwill 门店都能获得准确信息，也是正面回击长期以来围绕 Goodwill 品牌的各类谣言与误导性言论的关键战略举措。 - **网络规模与离散挑战**：全美 150 家独立分支网络拥有 150 个独立站点，内容冲突与数据割裂严重。 - **SEO 向 AEO 演进**：顺应零点击（Zero-Click）搜索趋势，确保品牌核心事实能被 AI 检索生成模型正确引用与沉淀。 - **数据净化与重构**：统一 Schema 结构化数据、重构 Q&A 问答生态，以单一权威事实源重塑品牌公信力与信任度。 --- ### [17:56](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1076s) AI 时代的组织重构：从具体执行者转向 AI 工作流训练师与调优专家 面对“在 AI 时代如何构建内部团队与外部代理商矩阵”这一让许多高管彻夜难眠的课题，Onney 与 Monica 分享了各自的组织变革实践。Onney 透露，Goodwill 营销部门在过去一年中完成了战略重组。她清醒地指出，在引入 AI 之前必须先梳理并修复内部原本低效或错误的运作流程；否则，直接将 AI 接入有缺陷的流程中只会导致“以更快的速度做错误的事”。在精简低效事项并明确核心业务优先级后，Goodwill 专门设立并正在招聘一个全新高阶岗位——“高级发展总监”（Senior Director of Advancement），该角色需具备深厚数字生态视野，专门负责用 AI 视角审视与改造部门的组织架构与业务系统。 Monica 则深入剖析了营销专家的角色演化。她认为，AI 工作流、智能体（Agents）以及底层平台的训练和搭建必须依赖具有深厚专业知识的业务专家，AI 本身无法凭空自我构建。但在未来，团队并不需要过去那么庞大规模的“纯执行型专家”。以往擅长 SEO 或 B2B 渠道营销的专家往往要将大部分精力耗费在繁琐的日常执行上，而现在这些执行环节已完全被 AI 智能体自动化接管。 ```markdown | 维度 | 传统营销团队角色 | AI 时代新型营销专家 | |:--- |:--- |:--- | | **工作重心** | 耗费大量时间进行手动文案、排版与渠道执行 | 搭建并配置 AI 业务工作流、引导智能体运作 | | **专业价值** | 个人经验驱动的单一技能交付 | 制定高质量交付标准、识别并剔除 AI 幻觉 | | **核心能力要求** | 传统 B2B 营销战术与流程熟练度 | 具备 AI 敏锐度，深度理解技术并善于精细化微调 | | **工具使用习惯** | 传统营销 SaaS 与手动办公软件 | 熟练协同 ChatGPT、Claude 等 AI 工具的主动创新者 | ``` Monica 表示，SOCi 并没有裁减团队编制，而是深刻改变了岗位职能模型。现在的营销专家核心任务是深度理解底层技术、精准界定高质量产出标准、进行模型与提示词微调，并敏锐识别潜在的 AI 幻觉与错误。在如今的招聘与内部晋升中，相较于只懂传统 B2B 营销的人选，她更青睐那些已经深度将 ChatGPT 或 Claude 等 AI 工具融入工作流程、具备主动创新与自动化思维的复合型人才。 ### [21:55](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1315s) 品牌的灵魂与摆脱低效重复劳动 在探讨营销团队的核心能力时，Goodwill 首席营销官 Onney Crawley 强调，无论技术如何演变，以人为本的叙事能力始终是品牌的护城河。Goodwill 的核心使命是帮助人们改变生活，这种具有温度的真实故事几乎构成了品牌独有的专有资产。她直言目前职场社交网络（如 LinkedIn）上充斥着大量缺乏灵魂的同质化生成内容（"this blop is sort of getting out of control"），因此营销人员必须具备敏锐的审美品味与判断力。AI 可以代劳繁杂的执行工作，但品牌的“灵魂”只能源于活生生的人；团队必须将精力集中在能让品牌展现鲜明观点的工作上，避免淹没在信息洪流之中。 高露洁-棕榄（Colgate-Palmolive）的营销与转型副总裁 Christy Borrowman 对此深表赞同，并生动指出了传统大企业内部极其痛苦的低效劳动现状：许多优秀营销人员每天花费海量时间，在数字资产管理系统（DAM）、产品信息系统（PIM）、演示文稿、创意素材与包装宣称等多达 35 个互不相通的独立系统中，机械地复制粘贴完全相同的文案。团队当前的目标正是构建智能协同系统来替代这些重复劳动，将专业人才的大脑从繁重无趣的琐事中解放出来（"let's get out of the crappy work"），让他们重新专注于真正具有战略价值的品牌建设、情感共鸣与叙事创作。 ```mermaid flowchart TD accTitle: AI 人力解放与价值升维流程 accDescription: 展示营销人员如何从重复性系统录入转向高价值品牌叙事 subgraph LowValue ["低价值机械劳动 (由 AI 与系统自动化接管)"] A["同一文案内容"] --> B["DAM / PIM 系统"] A --> C["包装宣称与文档"] A --> D["35+ 独立企业系统"] end subgraph HighValue ["高价值核心能力 (由专业营销人才主导)"] E["品牌灵魂与核心叙事"] F["敏锐的商业与创意判断力"] G["以人为本的情感共鸣"] end LowValue -.->|"系统自动化解绑"| HighValue ``` --- ### [24:04](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1444s) AI 时代的营销人才选拔与招聘思维转变 针对“企业是否正在改变营销人才招募策略”的问题，与会嘉宾分享了各自观察到的显著分化与策略调整。一方面，由于大量初级执行工作已被 AI 取代，传统招募刚毕业新人的路径一度受阻。而在重启实习与初级岗位后，团队发现应届生呈现两极分化：一部分年轻人心理上对 AI 产生排斥，误认为使用 AI 等同于“作弊”，需要管理者推着往前走；另一部分人则是 AI 原生代（AI-native），凭借对新工具的熟练驾驭，在工作效率和产出上甚至大幅超越资深员工。传统营销人员若固守过去 20 年的公关或传播旧套路，反而容易成为企业转型的包袱。因此，现在的招聘不再仅看垂直领域的履历年限，而是更加注重“思维范式”以及跨行业的借鉴能力。 Sequel.io 营销副总裁 Kathleen Booth 从实操层面提出了更苛刻的选拔标准：她不再满足于候选人口头上宣称“我喜欢使用 AI”，而是要求直接展示用 AI 亲手构建的具体成果（例如为自己求职量身定制的网站或互动问答 Bot）。此外，由于 Sequel.io 的业务拓展代表（BDR）团队直接向营销部汇报，并承担 100% 的销售线索管道（Pipeline）生成重任，初创公司在追求极致速度的压力下，根本无力花费大量时间从零培训无经验的新人。为此，她转向了创新的招聘孵化合作模式： - **合作机构与模式**：与专注于职业转型培训的机构 SV Academy 合作，由其对零销售经验的跨行业候选人进行电访（Cold Calling）与异议处理等高强度实战集训。 - **无风险试用机制**：候选人前 3 个月挂靠在培训机构，企业支付薪资且无需招聘中介费；若考核达标即可直接转正入职。 - **快速见效（Time to Impact）**：新员工入职第一周即可直接预约客户演示（Booking Demos），极大缩短了人才产生业务价值的周期。 - **内部人才激活**：高露洁团队则侧重于从内部挖掘具有好奇心与拥抱新技术意愿的早期采用者（Early Adopters），通过全员培训提升存量人才技能。 --- ### [29:01](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1741s) 从事后归因到前瞻预测：营销衡量的范式转移 在探讨数据与指标评估时，Christy Borrowman 指出传统营销衡量体系存在明显的“滞后性”（Retroactive）。行业长期依赖回顾性的营销组合模型（MMM），去分析 3 个月、1 年甚至 3 年前的营销表现与事后归因。然而，在 AI 赋能的当下，内容生成的高速度、高通量与高度个性化对评估机制提出了即时性要求，营销衡量正从“事后复盘”转向“近实时监控与前瞻预测”（Predictive Optimization）。 未来的评估体系需要先建立基准，并在活动投放过程中根据数据动态进行实时调优与策略转向。同时，营销人员必须摆脱对单纯点击率（Clicks）或传统广告投资回报率（ROAS）的机械式依赖。Christy 风趣地调侃了当下管理层常见的认知误区——高管们不能仅仅因为“自己在 ChatGPT 里提问，推荐的第一款牙膏不是我们品牌”就简单判定营销失败。营销团队必须建立更客观、更贴合当下消费者真实交互场景的全新度量衡。 | 评估维度 | 传统营销衡量模式 | AI 驱动的新型衡量模式 | |:--- |:--- |:--- | | **时间属性** | 滞后性、事后复盘（Retroactive） | 近实时、前瞻预测（Predictive / Real-time） | | **典型工具与指标** | 季度/年度 MMM 模型、点击量、传统 ROAS | 动态基准监控、跨渠道即时优化、受众交互预测 | | **执行机制** | 投放周期结束后复盘总结 | 活动投放过程中动态微调与资源倾斜 | | **决策依据** | 历史投放归因报表 | 数据流驱动的动态决策与算法优化 | --- ### [31:16](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1876s) 应对高管与董事会：前置指标与最终投资回报 针对“如何向管理层与董事会汇报 AI 投资价值”的问题，Kathleen Booth 结合其风投支持企业（Venture-backed）的背景剖析了沟通策略。随着资本环境从以往不计代价狂热追求增长（Growth at all costs）转向重视合理 ROI 与精细化运营，任何在 AI 工具或时间上的投入都必须受到严苛的财务审视。在衡量 AI 价值时，她明确区分了“前置指标”（Leading Indicators）与“后置指标”（Lagging Indicators）： - **前置指标（Leading Indicators）**：关注团队工作流程的效率提升。例如 Sequel.io 在内部推行“吃自家狗粮”（Dogfooding），将自身产品的 AI 功能深度应用于网络研讨会（Webinar）流程中，成功减少了 80% 的执行耗时。 - **高管与董事会关注的现实**：Kathleen 直言，尽管节省 80% 运营时间的前置指标令执行团队倍感振奋，但董事会根本不在乎内部工时节省（"the board does not care"）；管理层与投资人唯一关注的是这些效率提升是否最终转化为收入增长、成单转化与管线规模等硬性商业结果（后置指标）。 ### [32:38](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1958s) 董事会视角下的关键指标：从先行指标转向商业产出 对于软件和 SaaS 企业而言，年度经常性收入（ARR）始终是最核心的度量标准。在董事会层面，投资人们日益关注一个关键指标——“每位员工对应的 ARR”（ARR per employee）。董事会看到像 Cursor 和 Lovable 这样极具代表性的 AI 原生企业展现出极高的人效，即便传统或转型中的软件企业并非这类原生公司，但同样作为风险投资支持的企业，资本必然要求团队展现出同等维度的生产力飞跃。 这种高维度的衡量可以逐级向下拆解至具体业务线：例如单兵销售的产出效率（单个销售代表能带来多少营收增长），或是各渠道的投资回报率（在维持或微增预算的情况下，如何非等比例地扩大渠道营收贡献）。尽管随着全渠道触点的复杂化，归因分析正变得愈发困难，但向高管层及董事会汇报时，必须清晰区分内部过程指标与外部商业结果。 - **过程指标 vs 最终结果**：团队内部应当密切追踪过程性的先行指标（Leading Indicators），但面对董事会与 CEO 时，应重点呈现能够直接体现商业影响的产出指标（Impact Measurements）。 - **人效基准的重塑**：AI 原生公司极高的人效正在重塑资本市场对 SaaS 公司的估值与效率期待，销售人均产能与渠道 ROI 成为核心关注点。 --- ### [33:45](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2025s) 展望三年后：AI 对营销的深远重构与“品牌套利”的回归 讨论转向了对未来三年的前瞻：三年后回顾当下，AI 对营销影响最深远的将是什么？哪些工作会贬值，哪些又会变得更为宝贵？Christy Borrowman 引用了 Glean 公司 Rebecca 的一句话：“不要把工作的灵魂自动化掉”（Don't automate the soul out of work）。品牌建设的核心始终是让消费者建立情感认同、爱上品牌。如果把激发营销人员热情的核心工作完全剥离，品牌将失去灵魂与感知优势。AI 的使命应当是自动化与放大那些繁琐低效的杂务（"crappy stuff"），从而让“人机协同”（Human-in-the-loop）重新聚焦在真正重要的实质性工作上。 Onney Crawley 与 Kathleen Booth 也表达了高度共鸣。Onney 指出，正如当年电力赋能冰箱、互联网赋能银行业一样，AI 将逐渐退去概念光环，成为像水电一样无处不在的基础工作方式，但驱动品牌忠诚与情感共鸣的实质工作依旧需要人类来完成。Kathleen 进一步从“套利”（Arbitrage）的视角剖析了营销演进史：过去十年营销人依靠技术、自动化、提速和个性化获取竞争优势，但当这些工具全面普及成为行业标配（Table Stakes）后，技术套利的红利将彻底消失。营销将迎来钟摆式回归——重返《广告狂人》（Mad Men）时代的经典法则，将“品牌本身”作为 AI 无法复制的核心套利支点。 ```mermaid flowchart TD accTitle: 营销核心竞争优势的演进路径 subgraph Past["过去十年：技术驱动套利"] T1["流程自动化与提速"] T2["大规模个性化与工具红利"] end subgraph Present["当下转变：AI 成为行业基准（Table Stakes）"] AI["AI 深度融入工作底层<br/>效率工具高度普及并拉平差距"] end subgraph Future["未来三年：回归经典品牌套利"] B1["人类情感与品牌灵魂"] B2["差异化感知与忠诚度构建"] B3["《广告狂人》式的经典战略思维"] end Past --> Present Present --> Future ``` --- ### [37:43](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2263s) 快问快答：如果每周因 AI 节省 10 小时，你会做什么？ 主持人向全场嘉宾抛出了一个快问快答问题：如果 AI 能够每周为你省下 10 个小时，你会把这些时间用在哪里？各位高管从不同的战略维度给出了回答： - **聚焦实时竞争策略与真实营销**：摆脱日常琐碎任务后，产品营销（PMM）团队可以迅速响应市场动态，专注于当前正在发生的实时竞争打法与真正的市场战略。 - **深入一线倾听一线声音**：Goodwill 在全美有 150 个本土会员机构，CMO 希望将省下来的时间投入到与各地基层 CEO 的面对面交流中，深入了解一线挑战，从而制定更扎实的全国性战略。 - **给大脑留白与激发灵感**：在所有人都在为追赶 AI 学习曲线而疲惫不堪、焦虑过载的当下，营销团队需要空闲的时间（甚至“泡个热水澡”）来放松，重拾以往供员工自由探索灵感项目的机制，让真正的创意自然涌现。 - **五五开：一半探索工具，一半链接客户**：将 50% 的时间作为“AI 技术迷”去上手试玩每天层出不穷的新工具；另外 50% 的时间则全部花在与客户和买家深度交流上，挖掘他们真实的生活痛点与核心顾虑。 --- ### [40:24](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2424s) 现场互动：数据清洗治理与构建企业“第二大脑” 在现场提问环节，观众针对实际落地中的数据与架构痛点向嘉宾发问。关于“如何利用 AI 在不牺牲质量的前提下补全数据缺口并保持完整性”，嘉宾坦言全行业仍在积极探索分类分层与自动清洗方案，核心在于避免“垃圾进、垃圾出”，同时在局部场景（如本地市场智能）中，可通过统计学显著维度的信号数据拟合来平滑噪声。 随后，来自 Poff Labs 的观众 Matt 追问 Kathleen：在构建企业大脑（Company Brain）时，若不同数据源（如 CRM 记录与 CDP/客户数据平台）发生冲突，或面对非结构化、非逐字一致的客户反馈碎片时，系统应如何可靠地提炼洞察？ ```mermaid flowchart LR accTitle: 企业第二大脑数据处理与交互架构 subgraph Ingestion["底层数据与向量检索"] Raw["碎片化非结构化数据 / CRM / 反馈"] --> SB["Supabase 数据库"] SB --> Vec["Vector Search 语义向量检索<br/>（基于语义理解而非单纯字面匹配）"] end subgraph Output["交互与校验层"] Vec --> Slack["Slack 交互界面<br/>（拟人化交互 / 事实问答）"] Slack --> Conf["置信度反馈<br/>（展示百分比概率，不确定时主动提示）"] Conf --> HITL["人工介入校正<br/>（Human-in-the-loop 纠错与微调）"] end ``` Kathleen 结合自身实践分享了具体实现路径： - **透明的交互与置信度反馈**：她们在 Slack 内部署了交互界面。该助手具备拟人化的表达风格，当遇到不确定的信息或冲突时，它会主动说明并给出置信度概率（Confidence Probability），团队成员可以随时在群聊中进行人工介入与校偏。 - **引入向量检索（Vector Search）捕捉语境**：Kathleen 在其个人 Substack 专栏《Code Meets Creed》中详细记录了这一转变。她将原始非结构化数据存储于 Supabase 中并叠加向量检索能力。与传统的文本精确匹配（Exact Word Match）不同，向量检索通过理解语义和深层含义来查询数据，能够极大地捕捉到碎片化客户反馈中的细微语境（Nuance）与上下文，成为构建可靠企业知识库的关键转折点。 ### [43:36](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2616s) 依托“品牌大脑”校准发声与圆满结语 在探讨如何平衡用户声音与品牌定位时，讲者强调品牌必须清晰了解自身的内核与认知。虽然外界和用户的真实反馈弥足珍贵、堪称“真金白银”（"what people say is gold"），但品牌最终仍需掌控并传达属于自己的故事。因此，任何对外传播都必须时刻回溯并引用核心的“品牌大脑”（brand brain），借此明确品牌真正想要表达的内容，并据此对各类信息与叙事进行优先级排序。将这些机制与 AI 系统紧密连接其实十分顺畅且易于实现。 随着这一核心观点的总结，主持人提醒讨论时间已到，圆满宣布本场圆桌研讨会结束，并向现场所有观众与嘉宾致以诚挚的感谢。 - **品牌大脑（Brand Brain）的核心作用**：无论收集到多少用户反馈与外部素材，输出内容都必须与品牌自身的定位、核心叙事和优先级保持对齐。 - **研讨会收尾**：嘉宾们在热烈的掌声与致谢中结束了关于 AI 如何驱动营销效率与精准度的全场分享。 ## Notable quotes > [00:04](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=4s) 现场有多少人喜欢淘二手货（thrifting）？ — Onney Crawley > [05:48](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=348s) 感谢 TikTok，我现在每个月需要生产 10,000 条内容。 — Christy Borrowman > [21:55](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1315s) this blop is sort of getting out of control — Onney Crawley > [21:55](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1315s) let's get out of the crappy work — Christy Borrowman > [31:16](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=1876s) the board does not care — Kathleen Booth > [33:45](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2025s) Don't automate the soul out of work — Christy Borrowman (quoting Rebecca from Glean) > [43:36](https://www.youtube.com/watch?v=1KZsG08Ds1o&t=2616s) what people say is gold — Kathleen Booth ## Mentioned Marjorie McCarthy, ISS North America, Monica Ho, SOCi, Ford, Ace Hardware, Jersey Mike's, Onney Crawley, Goodwill Industries International, Christy Borrowman, Colgate-Palmolive, Kathleen Booth, Sequel.io, Datadog, Clay, Morningstar, TikTok, Pavilion, Cerebro, Slack, Intercom, SV Academy, Cursor, Lovable, Glean, Mad Men, Matt, Poff Labs, Supabase, Code Meets Creed ## Video description Marjorie McCarthy — Chief Marketing Officer, ISS North America Onney Crawley — Chief Marketing Officer, Goodwill Industries International Christy Borrowman — VP of Marketing & Transformation, Colgate-Palmolive Kathleen Booth — VP Marketing, Sequel.io Monica Ho — Chief Marketing Officer, SOCi AI is enabling marketing teams to do more with less—faster. This panel features leaders who are using AI to automate creative generation, tailor campaigns, and boost conversion rates with data-driven precision. We’ll explore how AI is changing the economics of customer acquisition and reshaping go-to-market strategies. You’ll also hear insights on managing data privacy, attribution challenges, and brand authenticity in a world of synthetic content. [Source](https://www.youtube.com/watch?v=1KZsG08Ds1o)

[0:04] Onney Crawley: 现场有多少人喜欢淘二手货（thrifting）？ [5:48] Christy Borrowman: 感谢 TikTok，我现在每个月需要生产 10,000 条内容。 [21:55] Onney Crawley: this blop is sort of getting out of control [21:55] Christy Borrowman: let's get out of the crappy work [31:16] Kathleen Booth: the board does not care [33:45] Christy Borrowman (quoting Rebecca from Glean): Don't automate the soul out of work [43:36] Kathleen Booth: what people say is gold

[5:48] Christy Borrowman: Thank you TikTok, I now need to produce 10,000 pieces of content a month [14:40] Monica Ho: The upfront lift on building actual AI workflow and modules, it's real... you think it's so simple, you're like, 'Oh, I'm just going to prompt this. I'm going to feed it a couple of things and then I'm going to get what I need out of it.' And you would be surprised. You don't get what you thought out of it. [17:56] Onney Crawley: Knowing that we wanted to bring in more AI capability, more AI learning, we had to sort of clean up our structure first because otherwise what's going to happen is we're doing the wrong thing... and you're going to bring in AI and just do that all better. [21:55] Christy Borrowman: I have marketers today who spend time putting in the same copy point in a DAM, a PIM, a binder, a creative asset packaging claim. It's all the same. And they're spending time doing that repeatedly because it's 35 different systems... let's get out of the crappy work and let AI do that. [29:01] Christy Borrowman: It can't be your division president sitting in the room saying, 'Well, I used ChatGPT and the best toothpaste is not coming up as us.' We need to get a lot better at the realities of what it takes to get there. But the future of measurement is absolutely much more, in my opinion, proactive versus the reactive. [33:45] Christy Borrowman: Don't automate the soul out of work.

[0:04] Hi everyone. Welcome to Scaling Smart. Um,  very very excited to have y'all here. Thank [0:11] you for getting up somewhat early to to come  join our session. I'm Marjorie McCarthy. I'm [0:16] chief marketing officer for ISS. If you are not  familiar with ISS yet, we are a global workplace [0:24] experience and facility management services  company. We have over 320,000 employees working [0:33] across almost 60 countries around the world,  all delivering hospitality in the workplace. So, [0:39] we are essentially the people that power all of  the spaces where we work and have experiences like [0:47] this one. Um, I have the privilege of leading our  marketing and communications team in North America [0:54] and I also have the privilege of moderating our  panel today. And we have a really fabulous group [1:02] of marketers here to talk about this very  interesting topic. And I'm going to ask each [1:07] one of y'all to introduce yourselves. Please tell  us a little bit about you, your role, your team, [1:14] your remitt within your organization. Um, I'm  just going to go down the line here. Monica, [1:19] you want to start? Sure. Hello everyone. Monica  Hoe, um, CMO of Soshi. Soshi is an agentic [1:26] marketing platform specifically built for brands  that don't just operate out of one location or [1:33] one storefront, but hundreds to thousands of them.  So, think Ford Motor Company, Ace Hardware, Jersey [1:39] Mike. Those are some of our our clients. Um, and  what we do is we have local brand trained agents [1:45] that do the critical local marketing work at each  location that is essential to being found in AI [1:51] search, traditional search and social media sites.  So, um, think of like local reviews or optimizing [1:57] your local pages, things like that. Um, as the CMO  of SOI, I run all of marketing. I've been in the [2:04] seat for eight years, so have seen a tremendous  amount of transformation. Um, what we're really [2:09] focused on, uh, in our team right now is really  paying off what we promise to our customers, [2:15] which is that Agentic AI can really transform  the way that marketing gets done. All right, [2:23] good morning. How many um, thrifters do we have  in the room? Can I see show? I love this. Anie [2:29] Cwley. I'm the chief marketing officer of Goodwill  Industries International. We are a federated model [2:35] actually made up of 150 different organizations  who are deeply embedded in our local communities. [2:40] And all of you probably know our thrift stores.  You probably know about our donation services. [2:46] What you may know less about is that the reason we  actually exist and why we take your donations and [2:52] why we sell them is we use that revenue to provide  free services in those local markets for people [2:57] who are looking for job services and training.  So, we're all about workforce development and [3:02] economic empowerment. And a big part of my remmit  in this job is to make sure that those stories [3:06] are getting told and that we're getting in more  donations. We're getting more people to come and [3:10] thrift so that we can actually serve more people  in local communities. And because we are federated [3:15] and everyone's kind of doing their own thing in  the local market, I also need to make sure we [3:19] are still moving as one brand. And I think AI is  going to go a long way in helping us do that even [3:24] better and faster than we have in the past. Nice.  I'm inspired already and I was already a fan. So, [3:32] uh my name is Christy Borrowman. I work  for Colgate Palm Olive. Uh I have for the [3:37] last uh 11 years and have made my career on uh  transformation um really it started in digital [3:46] data media those things and then now my current  role as of seriously this month AI transformation [3:55] as part of the the marketing arm right so we are  looking to move end to end how how are we going to [4:00] do this what's the future going to look like for  our organization I've also done time at the agency [4:06] see and and work really closely with many of  our developers and and I think that that's [4:10] also going to come in to play um connections for  the future. So happy to be here. Thank you. Hi [4:16] everyone. I'm Kathleen Booth. I am the VP of  marketing at SQL.io. Uh SQL it was the first [4:23] webinar platform purpose-built to host webinars  on your website. Uh today we are evolving into [4:29] really an agentic go to market platform because  of the way we embed we collect a unique set of [4:34] first-party data encompassing all of the webinar  data as well as all the website page visit data [4:39] before during and after the webinar and we're able  to combine that and score it at the the contact [4:44] and the account level and companies that work with  us and use our platform are there companies like [4:50] data dog clay morning star they're using that  data to power their broader gotomarket motion [4:57] Um, and so we we're all about basically helping  these companies use the agentic front end that [5:03] we're building and use AI to scale the impact  of their video and their webinars. Specifically, [5:09] um, my remitt is really around balancing brand  uh, and a human first kind of approach to go to [5:16] market with AI. We are a an early stage company.  uh this is kind of what I've built my career on is [5:23] working in startups and early uh scaleups and  we are looking to grow in the same way that a [5:28] lot of the AI native companies are growing. So  the remmit I have is headcount is not going to [5:33] expand proportionate to revenue and uh we're  really leaning in hard on AI to help us make [5:39] that happen but without sacrificing brand as we  go. Thank you. Um okay so first question starting [5:48] kind of at the 30,000 foot level and this is  for for everyone where do you see AI driving the [5:56] greatest competitive advantage for marketing today  in this moment and anybody can start uh I can I [6:04] can start if you want um it's it's two things for  for me it's around data uh if you don't have your [6:12] data foundation in order um then you can't really  use AI to augment everything on top of that. So [6:20] um really helping us build that up. Obviously  we have to get it organized. We have to get the [6:24] taxonomy right. We have to connect the dots. But  AI I think can then sit on top of it and I believe [6:30] empower more people to actually better utilize  the data for what we really really want to do. [6:37] The other piece are the basics of content, right?  Um the scalability, the velocity, the volume. We, [6:45] you know, thank you Tik Tok, I now need to  produce 10,000 pieces of content a month. Um and [6:51] so regardless of how many people you have on your  team, you you never have enough time to get that [6:55] done. So AI will make a huge transformation or  is um already impacting the the opportunity to do [7:01] that. I was just say I'll jump in because I have  almost the exact same answer. um and just add a [7:07] little flavor to it. Um the data piece I I cannot  underscore enough and I think what to me really [7:13] is going to unlock competitive advantage is the  tribal knowledge that we have within our companies [7:18] for the first time. We have an opportunity to  harness it and a great example is you know we [7:22] just built a company brain which I I'll share more  about in a minute but it includes not just the [7:26] data that are in our that's in our go to market  uh platforms but also call recording information. [7:33] um you know we can now search that and and pull  out spoken knowledge that I think otherwise gets [7:39] lost without AI. So that to me is huge. And the  second piece I I feel like as a marketer the [7:45] the biggest problem I've had and I've seen before  joining SQL I led marketing at Pavilion which is [7:50] the largest global private community of gotomarket  executives in the world. We had 10,000 members. [7:56] What I saw across that community uniformly was  an issue where marketers poured so much energy [8:01] into creating content and not nearly enough energy  into distributing it and and promoting it. And AI [8:08] for the first time really gives us an unlock to  solve for that which I just think is incredibly [8:13] exciting. Al all so true. I couldn't agree more  with both of those things. And I will sort of add [8:19] on to that two things. One is that I think the  tribal knowledge aspect just the democratization [8:23] of data across an organization. We all know the  situation, right? You've done this work. You've [8:28] done this research and it's now sitting in your  file in your folder. Maybe you try to centralize [8:32] it through your team, but the organization doesn't  have broad access to it. And so using AI to really [8:37] help give everyone the same information at the  same time to make the best decisions for the [8:42] business. The other piece is, and this is really  much more consumer centered and consumer focused, [8:46] the ability to really grow and adapt with  your consumer in real time. As a marketer, [8:51] one of my biggest sort of I guess pet peeves is  that we do the segmentation work, right? You spend [8:56] all this money and time and do this big research  body of stuff and now it's sitting there like, [9:01] okay, now I know who my audience is and  what they think and what they do and how [9:03] they behave and that stuff gets dated really  quickly, especially today. So now having the [9:09] ability to in real time adapt and grow and learn  with your audience so you can tightly target them [9:15] um and tightly understand how to engage them,  I think is really really going to be valuable. [9:20] surprise. I'm going to say data also, but I have  a little bit of a different response. So, where I [9:26] sit at, we're a marketing platform. What we love  is marketers that know how to leverage their own [9:31] data, which is incredible, but the reality is  that you have only so much of that data. And as [9:36] on said, like the marketplace is changing so much.  Search is changing. What we do is we're one of the [9:43] platforms that can bring in local data signals and  actually pair that with your own location data, [9:50] your own um business data. Why that's important  is because you might know a little bit something [9:55] about your different businesses, the products  that they offer, but do you know how they're being [10:00] searched in that market? Do you know exactly what  your consumers like or dislike about that specific [10:06] business? Do you know what the competitive set  looks like in that local market? that's changing [10:10] all the time. Those are the data signals that  you now can access with really good AI data [10:16] platforms and then merge that with your data. The  cautionary tale I would tell you is that there is [10:21] a lot of access to data. Everybody's got an MCP  and everybody wants to hook it up and you want [10:25] to start leveraging that data. Think about is that  really what you want to do. Do you really want to [10:31] build a a marketing platform or an application?  There's a lot of really great MCPs out there, [10:37] but you got to be careful that the data is  complete and that there's enough of it so you're [10:41] not feeding something to your own systems that's  going to allow the AI to hallucinate. Okay, all [10:47] those responses got me very excited for this next  question, which is let's get a little bit more [10:52] tactical. can't would love to hear some specific  examples of um a tactic or a use of AI that y'all [11:02] have recently found in your organizations or  you're in the process of deploying. Um Kathleen, [11:09] I think we're going to start with you on that  one. Sure. So, we just built we had a company [11:14] offsite recently and we had a little hackathon  and I was part of a team that built something I [11:21] am so excited about. Um, the nerd in me is really  coming out now. But we built what we call Cerebro. [11:28] I'm an X-Men fan. If anybody here likes X-Men, you  know, Cerebro is that big Yeah. the big the big uh [11:35] brain that can see all the X-Men everywhere in the  universe. So, we wanted to build a company brain. [11:41] And what we did is we listed every source of  data that we had in the company. Everything from [11:47] our CRM, our marketing automation, our product  usage platform, our customer support platform, [11:54] our call recording for sales, for CS, um our  project management, even our employ uh internal [12:01] HR docs, like literally anything you might have  a question about if you work at our company. And [12:07] we we piped it in to this system and the front  end of it actually lives in our Slack instance. [12:15] And so now anyone can go in and either in the  shared Cerebro channel or even in a private DM [12:20] if you don't want to share what you're asking, you  can ask it a question. And this goes back to the [12:24] point about data. What is unbelievable about this  thing is that it can combine data from different [12:30] sources to create really contextual answers.  So I had a field day playing with this when [12:36] we first built it and asking it things like um  what how do our customers talk about our product [12:43] in their own words? what are the words they use?  You can imagine as a marketer like gold in terms [12:48] of messaging. And so it looked across, you know,  CRM notes. It looked in all of our call recordings [12:54] which span two different platforms. It looked in  our intercom help chats and it came back and it [12:59] stack ranked the most common words and phrases our  customers used to talk about us. I asked it about [13:05] uh win loss analysis. Um, I asked it about  competitive positioning and if we, you know, [13:12] what do customers love most about us VisaV  competitors? I mean, the number of use cases [13:18] for this is insane. But here's what's really  cool. That's just how I'm using it. Our CS team [13:22] is actually using it now in place of our CS uh,  platform because it's providing better answers. [13:30] It sees our help docs. It sees all of our existing  help chats. And so while they're on a support call [13:36] with a customer, they can go in and ask it a very  detailed question and it will come back, answer [13:42] it in the thread. And what I love is we all have  visibility. So for example, our CTO very commonly [13:48] will jump in and if there's some misinformation  or it's an incomplete answer, he'll respond. And [13:53] the way it's set up in in um Slack is it will then  it has recursive learning, which in layman's terms [13:59] means it learns from the answers it gets. You can  have a conversation with it. it will write back to [14:04] its memory and get smarter every time we correct  it. So that to me is just like the dream state. [14:11] I I feel like I barely even tapped into how we're  going to use it, but it's super exciting. Just to [14:16] comment on that, that actually in a couple of  the round tables I was in yesterday that came [14:20] up a lot that if you build the brain, right, what  happens when it starts perpetuating inaccuracies, [14:28] right? And and what do you do about that? So I  think that's great that you guys built that in. [14:32] I have about 10 more questions about Cerebro. Um  Monica, I'm gonna ask you to share an example now, [14:40] please. Sure. Um I'll get a little bit more  specific. I think um one of the things that we've [14:46] been challenged with, especially as a technology  company is how quickly our product team can now [14:52] ship products, new products and enhancements.  And so we found ourselves in a situation where [14:58] marketing became the bottleneck, not product this  time. And so we were being, you know, it typically [15:04] takes my team if it's a, you know, a brand new  offering to the market about 8 to 12 weeks, [15:08] right? You've got PMM, you've got enablement,  you all have all these teams. Um, we were able [15:13] to build modules that allow us to take a lot of  that enablement work, a lot of that content. Um, [15:20] we had to build these modules. So, uh, the upfront  lift on building actual AI workflow and modules, [15:27] it it's real. like it took us months to do it  because you think it's so simple, you're like, [15:31] "Oh, I'm just going to prompt this. I'm going to  feed it a couple of things and then I'm going to [15:34] get what I need out of it." And you would be  surprised. You don't get what you thought out [15:38] of it. So, after some fine-tuning, a couple  of months, we now have these modules. We were [15:43] able to take our go to market from 8 to 12 weeks  down to 1 to two weeks, which is incredible. So, [15:50] um it's just a really great unlock if you're  willing to invest the time up front. Again, [15:54] I think the worst thing is thinking you're just  going to feed it a few prompts and it's going [15:57] to give you what you need. You need to you need  to watch it. You need you need human oversight, [16:02] but it's been a scale uh and a game changer for  Soshi. That's awesome. Um Anie, talk to us a [16:10] little bit about some work y'all have been doing.  Yeah, so we we so I mentioned earlier we have [16:15] 150 different organizations. That means we also  have 150 different websites. So, good times. Um, [16:23] and what we're trying to focus on now is making  sure that we are discoverable and that we are [16:27] consistent and that we have awareness and depth  and clarity with our content. When you think about [16:32] SEO and how things are migrating to now more of  sort of an AEO, right? We're not no longer waiting [16:37] for links. You're getting this whole AI summary  and you don't even go to the website, you get [16:40] what you need. How do we show up? And because we  have so many different websites, there's so many [16:45] different pieces of content that may contradict  each other. Um, they don't corroborate each other. [16:51] it's not cited. And so, you know, AI says, "This  is incredible. I'm not going to use this." And so, [16:56] we don't show up. So, the work we're doing right  now is to really kind of scan and crawl all of [17:01] these sites to understand what needs to change.  How do we change the schema, the back links, [17:06] the setup, the Q&A, all the different things to  allow us to show up as sort of one brand. So, [17:10] regardless of where you are or what Goodwill  uh location you're looking at or looking for, [17:15] you're getting sort of one set of really clear  and cleansed data about the brand. And it's been [17:21] a pretty big undertaking, but I think we're really  excited about it because it's another way to just [17:24] combat some of the um I'll call them untruths  about Goodwill brand. I'm sure some of you have [17:29] heard some of them. Um so this is going to be a  big project for us and it's going to take the um [17:34] collaboration of all of our network organizations  as well. So we've started engaging all of them and [17:38] I think that's going to be a big part of how this  brand how we drive uh trust and truth among our [17:44] customers. Thank you. Um, okay. So, let's let's  talk a little bit about how we are thinking about [17:56] our teams in the age of AI. So, I know this  is actually this is a question that keeps me [18:01] up at night, and I'm not kidding. Like, what how  do I make sure that I've got not only the right [18:06] in-house team, but the right agency stack as as we  think about um being in the age of AI. Um, Anie, [18:14] you recently went through a team restructure  where y'all did some thinking about this. Will [18:18] you talk to us a little bit about that? Yes, I  uh recently went through a process last year uh [18:24] for the purpose of doing a strategic realignment  of my department. Um, two things. One, we knew [18:31] that we needed to move faster um just given the  categories that we play in and making sure that [18:36] we can meet the moment of our different audiences.  I also knew that we had just some intrinsic issues [18:42] within the team and how we operate. And so knowing  that we wanted to bring in more AI capability, [18:47] more AI learning, we had to sort of clean up our  structure first because otherwise what's going to [18:53] happen is we're doing the wrong thing um and doing  things the wrong way and you're going to bring in [18:57] AI and just do that all better. Um and so it was  a real focused process of understanding like what [19:02] is the work where do we need to focus what are  our priorities and what do we sort of streamline [19:08] um in terms of what might be able to be automated  or what just needs to be what needs what might [19:12] not need to be done at all. So with that  restructure we're actually bringing in a new [19:16] role I'm currently in process of hiring that role  right now someone who sort of understands sort of [19:21] the digital landscape because a big part of our  path forward is really driving stronger digital [19:26] ecosystems and stronger communities. someone who  understands that landscape and understands how [19:31] to look at our structure and our systems through  an AI lens and that will really kind of be a part [19:37] of how the entire department sort of shifts  um and adapts to to what you know the markets [19:42] are demanding today. Sounds like a very cool  role. Don't know if there's any uh out there. [19:47] So there any senior directors of advancement in  the in the in the making? Let me know. Uh Monica, [19:55] you're you're obviously at the forefront of  this conversation, both advising clients, but [20:00] also thinking about it, you know, with regard to  your own team. Talk to us a little bit about that. [20:05] Yeah, I think um when I think about marketing  teams, um and we talked a little bit about, [20:11] you know, the go to market process I just rebuilt.  Actually training your AI workflow, AI agents, [20:19] a platform, it actually takes a human, right?  you're not going to get that done with AI. So, [20:25] I think marketing experts are still going to be um  a very necessary um role within marketing teams, [20:32] but the the reality is you're not going to need  as many experts, right? I think in a lot of our [20:37] teams, we've had a lot of experts, and those  experts might be great in SEO, they might be [20:41] great in B2B channel marketing, um but they also  had to do a lot of execution. Now with AI, you can [20:47] actually get rid of a lot of that execution work  and give that automate that through an agent. Um, [20:53] what that does and what we've been changing  on our team is that not necessarily do you [20:58] need less less staff, you actually need your  staff doing different things. So if you have [21:02] an AI workflow built and you have an agent  that's actually executing tasks, you don't [21:07] need that person doing that. You need that person  to be very good at understanding the technology, [21:13] how to fine-tune it, and to be expert at what that  output should look like and fine-tuning and making [21:18] sure that output is what it should be, right?  Identifying hallucinations and other things. [21:23] And so when I look at my staff today, as we, you  know, we haven't reduced our staff at all. I think [21:28] we've really changed up some of the positions. But  I will tell you that as I'm looking for new staff, [21:34] if I have somebody that's left or we need to  um promote people, I am looking for people that [21:40] um are not necessarily just great at B2B  marketing. I'm actually looking for people [21:45] that are very um they're already innovating  with AI, even a chat GBT or a claude co-worker [21:55] and I'm looking for proof. It's that type  of thinking um I think is kind of it's going [22:01] to be kind of the the the bare minimum for  the marketing team of the future. In fact, [22:05] can I just add to that quickly because on my team  a big part of Goodwill is the stories that we tell [22:11] and I always say we are we are a very human um  centered organization. We are in the business of [22:17] helping humans and so the way that we tell those  stories is really really important and almost [22:22] proprietary to the Goodwill brand. I've seen a lot  out there on social media. We all have LinkedIn [22:27] is starting to turn me like I I mean I can't this  blop is sort of getting out of control. Um and so [22:34] having people who understand how to tell those  stories who can look you know with sort of a a [22:39] lens of good judgment so you you you can get  people to do sort of some of the executional [22:44] stuff but there's also sort of this there's a  soul that your brand has and that's going to [22:49] come from the people. So understanding where to  place them and to your point it's not you know [22:53] they don't have work they're doing different work.  So what is that work to make sure that your brand [22:58] still stands out with a clear point of view in a  proprietary way that doesn't just get lost in in [23:05] you know everything that's out there. I have an  add-on as well. What I what we're hoping with our [23:10] teams as well is I have marketers today who spend  time putting in the same copy point in a dam, [23:18] a py, a binder, a creative asset packaging claim.  It's all the same. And they're spending time doing [23:27] that repeatedly because it's 35 different  systems that you have to do that. We're [23:33] we're trying to get to the point where we're  augmenting that person, that brain. You know, [23:38] there's a whole system that we're working on.  Then they can actually get back to real marketing, [23:43] right? So, I actually do believe I need experts  to build brands. Maybe not in the B2B like you [23:48] said as an example, but I I do need people  to still know the heart, the storytelling, [23:53] the the actual soul of of the work there. So  that that's what we're kind of hoping to is [23:58] is the future as well. Like let's get out of  the crappy work and let AI do that. Just a [24:04] quick follow-up question to all of this.  Are you finding, and this is for anyone, [24:09] are you finding that you're looking where are  have you changed where you are looking for talent? [24:20] That's a great question. We probably should  leave. You know what's interesting? So, [24:25] I don't know if you guys are the same, but with  AI, I think a lot of the lower level work has [24:31] gotten eaten up a little bit. And so we had an  issue over the last two years where we weren't [24:35] bringing young marketers into the organization.  And we decided as a team this past year that we [24:42] were going to change that. We we started an intern  program and then we we opened up we took a senior [24:47] role and we opened up a few junior roles. What  was interesting and you it's it's really I can't [24:55] put my finger on it but some of the young people  they actually do not want to use AI coming out of [25:01] college. they feel like it's cheating and I've had  to really like pull them along. But then I would [25:06] say on the other hand, we've hired some um some  of our interns came in and just because they're so [25:12] native with a lot of the tools um they're kicking  my team's ass on things like I mean even though [25:19] they don't have the experience because honestly  you guys some and I'm going to say something that [25:24] might be a little controversial but marketers we  have a lot of you know we we build our careers up [25:29] and some of us like if you're in PR right and  you've done communications the way that you've [25:34] done communications for 20 years this is not  the way you need to be doing them now and it [25:37] actually hurts us. And so knowing where you need  to change those rules out and freshen things up [25:43] is really important. You I don't it's a good  question but I don't know if I'm looking in a [25:48] different place. I'm looking for demonstration of  a different mindset almost which I'm seeing with [25:54] certain marketers. Um, but also too, and I've  always done this, looking for people outside my [25:59] category who are bringing kind of a different  mindset because oftentimes, as we all know, [26:03] you can get great learnings about your business  and your category from outside the category. So, [26:07] it's really trying to find someone who's kind of  thinking sort of in, you know, in in the way that [26:12] that that needs to happen today. I'll just add  really quickly to that. Um, I I think for the most [26:16] part, I would agree I'm changing how I'm looking  and what I'm I'm demanding of candidates. Like, [26:20] I want them to show me their work. I don't want  to just hear I use AI. I like to use AI. I'm like, [26:25] "Show me something you built with it."  Even better, if you're looking for a job, [26:29] do something as part of your the way you present  yourself. Like, build yourself a website. I've [26:33] seen people build chat bots on their site that  talk about them and answer questions. But there [26:38] is one way I am changing where I look. Um, so I  don't know if anybody else here um as a marketer [26:45] has the I have the BDR team that rolls up to  me. So, essentially my team is responsible for [26:51] generating 100% of the company's pipeline. So, our  BDRs do all of our outbound. And the thing that [26:58] I've noticed that's really important for me right  now because we are a small and growing company is [27:03] is speed is essential. And I cannot afford to hire  BD a BDR who has no experience. Um, and what I [27:12] mean by that is usually BDR is the entry-le sales  role. Usually you get someone who's never worked [27:16] in sales. You're training them, teaching them  how to cold call. you know, they're doing the [27:20] work honestly that no one else wants to do in  sales, but it's important. Um, what I've done [27:26] is I I don't have the bandwidth to train a bunch  of new BDRs. We're moving too quickly. And so, [27:31] we actually have a relationship Hello. We have a  relationship with a company called SV Academy and [27:38] they uh take career switchers and people who've  never worked in sales and they put them through [27:43] a boot camp and they train them on the basics of  cold calling and objection handling and all the [27:48] things that I would have to do myself if I hired  and experienced BDRs and you get your candidate [27:55] from SV Academy. They actually work for SV Academy  for the first 3 months. And if you like the work [28:00] they're doing at the end of 3 months, it's a great  you they just become part of your team. There's no [28:05] finders fee, recruitment fee. You're paying their  salary while they work for SV, but it's like a [28:10] risk-free way of getting a pre-trained team. And  one of my BDRs might be in the room right now. Um, [28:17] she started with us and she was booking demos in  her first week. That is really rare when you hire [28:22] somebody who hasn't had the experience. And so I  think like to broaden that the point behind that, [28:28] I do think we're going to start looking for ways  to find people who have more of that baseline [28:33] experience so that we can have impact faster  because that's what AI really allows us to do is [28:38] get time to impact shorter. I would I would just  add we're probably looking more internally, right? [28:45] As you said, it's the mindset. Who are the people  who have been early adopters, who are curious, [28:50] who are doing those things. So we and we want to  augment the people that we've got. So we spend a [28:55] lot of time on the training, the implementation  and and that's actually made a huge difference. [29:01] Um okay, this would not be a marketing panel if we  did not talk about data and measurement. So as AI [29:10] becomes more embedded in marketing, how should  we be rethinking how we measure its impact for [29:17] marketing purposes? So Christie, you are obviously  thinking about this all day every day and your [29:24] role. Talk to us a little bit about that. Yes, all  day every day. And and it's a passion. So that's [29:29] a good thing, right? Um I actually was thinking  about this and um instead of getting granular, I'm [29:36] I am going to go with a theme today. Much of the  measurement that we do is very retroactive, right? [29:42] It's an MM from 3 months, three years, one year,  you know, it's a marketing mix. What has happened? [29:50] How did that work? And of course, we still need  those things. However, I think we're getting to [29:54] the point where we're bu building real time or  more near real time that's actually allowing us [30:01] to do much more optimization, which means the  future of measurement has to look different, [30:06] right? It has to be maybe what's our baseline?  um X looks good, but then within the campaign [30:14] if it's not optimizing in real time because  remember we still have the volume, velocity, [30:19] all of this personalization that has to be done  as well. I I think the future of measurement for [30:25] us is actually going to get a little bit more of  what's the predictive and how are we hitting the [30:29] numbers and then do you shift gears um more more  heavily, right? I think also over time if you're [30:35] still a marketer focusing on those you know basic  numbers the rorowaz and the clicks and you know [30:40] you can you can speak much more to where we need  to show up in the world where the people are the [30:46] measurement has got to change on what's successful  and it can't be your division president sitting in [30:52] the room saying well I use chat GPT and the best  toothpaste is not coming up is us right so that's [30:57] a measurement I you know we need to get a lot  better at the realities of of what it takes to [31:02] to get there. But um the the future of measurement  is absolutely much more in my opinion proactive [31:08] versus the the reactive. And just to follow up  to that are again as marketers, one of the things [31:16] we're always thinking about is okay, how do I  talk to my leadership? How do I talk to my seuite [31:21] about, you know, whatever aspect of measurement?  Um, Kathleen, talk to us a little bit about how [31:28] you're how you're having those conversations about  AI with your with your leadership in your seauite. [31:34] Yeah, absolutely. So, we're venture-backed  and there is tremendous pressure to hit [31:38] very aggressive growth targets, but the venture  landscape, as many of you are probably familiar, [31:42] has changed a lot in the last few years, and it  we've moved from this like growth at all costs, [31:47] just spend, spend, spend, and don't worry about  ROI to a little bit more of a of a reasoned [31:53] approach. And so I do have to justify expense uh  into AI. I have to also justify any time we're [32:01] spending on it. Um so the way I think about it is  there's leading and lagging indicators. Leading [32:06] indicators are what I think most people are  talking about right now. Um you know we might look [32:10] at jobs to be done within our teams and we might  say wow we saved all this time. Like for example, [32:15] we're dog fooding our own platform and we're  estimating that we're saving 80% of the time [32:20] it used to take us to execute a webinar program  by just using our own AI that's built into our [32:25] solution. That's a great leading indicator. Um  but the board does not care. I was going to say [32:32] they do not care. Uh it it's great. Our team  loves it. Um what the board cares about in our [32:38] case because we are software we are measured on  annual recurring revenue. And so one metric that [32:43] I see coming up a lot in board meetings is ARR per  employee. They want that number to get much higher [32:49] because they're seeing it in these AI native  companies like the cursors and the lovables of the [32:54] world. And you know it doesn't matter that we're  not cursor or lovable. We're still venturebacked. [32:58] And if they're allocating their capital towards  us, we need to show that we can have that same [33:03] kind of an impact. And I think you can take that  and you can take it down a level and you can look [33:07] at productivity per sales rep. you know, how much  revenue is one individual sales rep able to bring [33:13] in and can we increase that? You can look at,  to Christiey's point, ROI by channel, you know, [33:18] how are we increasing the amount of revenue we're  bringing in per channel, you know, while keeping [33:22] the spend steady or or like growing it but not  necessarily proportionately. So, I think we have [33:28] to think of it that way like yes, let's as an  attribution is getting more complicated. Oh [33:33] yes. Um, but yeah, so like let's internally look  at the leading indicators, but don't go to your [33:38] board or your CEO with that. Go to them with the  impact measurements. That's super helpful. Thank [33:45] you. Okay, so zooming a little bit back out again,  if we're sitting here 3 years from now, what will [33:53] we say was the way that AI impacted marketing the  most? How will marketing have changed the most? [34:02] what will what will be less valuable? What what  are we doing today that will be less valuable now [34:08] versus what will be more valuable? Um Christie,  I want to start with you on that one. Sure. So [34:14] um actually in a session I was in yesterday uh  Rebecca from Glean had a great quote that was [34:21] uh don't automate the soul out of work, right?  Which has has again summarized so many things that [34:28] are running around in in my head this week. And  I'm using that reference because I think looking [34:35] three years from now, we still have to build  brands. We still need people to fall in love with [34:41] us. And if we take away the work that actually  fills you, that actually brings you into the [34:48] office as a marketer or wherever work from home um  wherever you are, right? You you need that work to [34:54] matter. And and I believe the difference also will  show through um especially for brands like ours, [35:00] right? We're really proud of our science. We're  really proud of the perceivable superiority of of [35:06] how the product feels, looks, performs. And if you  take the soul out of the work to truly get there, [35:12] then you you've got nothing, right? So my hope is  that's why we can automate, augment, scale some of [35:19] the crappy stuff so that we can actually get back  to the real work that that matters from humans in [35:24] the loop um part of the process. Kathleen, Anie,  you want to add to that? I was just going to I [35:30] love what you said cuz I think it's so absolutely  true. Markers will still be here. The work is [35:35] still going to be here. I think right now because  we're sort of in this age of AI growth, there's [35:40] a lot of like AI enabled and AI powered which is  absolutely true but we're not sitting here saying [35:46] oh you know electricity powered refrigerator right  we're not saying you know internet powered banking [35:52] so it will be a leveler and it will sort of become  how work happens but the work is still there to be [35:58] there to be done right so when the internet came  we didn't all all of a sudden we weren't all of a [36:02] job not doing something so it's going to change  I think what you were saying it will change how [36:06] we work, but the work is still there. And as a  brand that drives in consumers, engages them, [36:13] um, and drives brand love, which is what we want,  and brand loyalty, you still need those humans. [36:17] That's who is going to get it done. I love that.  Kathleen, did you have something you wanted to [36:22] add there? Yeah, sure. Um, I think marketing has  always been in some respects a game of arbitrage [36:28] where you're looking for an edge, right? And in  the last 10 years, a lot of that edge has come [36:34] from technology and ways we can automate and move  faster and personalize and and I would totally [36:41] agree with what Anie just said. I think that is  all now kind of table stakes and and it's it's [36:48] going to eliminate those aspects as edges or as  arbitrage uh levers that we can pull which leads [36:56] me to I actually think AI is going to be one of  the greatest things that happened to marketing. [37:00] This might be controversial or maybe not  given that this is an AI conference, but [37:05] um but I think what it's going to leave us with  is brand as the main element of arbitrage. It is [37:13] the thing that AI cannot replicate and it is the  thing that the marketer is uniquely positioned to [37:20] do best. And so it's it's going to be a return  to some of the classic tenets of marketing that [37:27] maybe we've shuttered shuffled off to the side in  the last several years as we've been so focused on [37:32] exploiting this technological edge. Um we're going  back to a time of I call it back to the madmen [37:38] era, right? It's kind of like that pendulum swing,  right? It's going to make its way back. Yeah. [37:43] All right. Um, one last quick fire question  for everyone and I think we have maybe we'll [37:48] have like maybe time for one or two questions  from the audience if folks have questions. So, [37:53] start thinking. Um, let's assume that AI is giving  you 10 hours a week back. What are you doing with [38:02] those 10 hours? I Everybody gets 30 seconds on  that. I would say spending There's so many timely [38:09] things that are happening right now, right? and  our teams have been just buried in tasks. Uh when [38:15] I when I talk about the the go to market process  that we just um made very fast, my PMM team now [38:21] is focused on competitive plays that are happening  right now because we've learned something new in [38:26] the market. So I think going back to again real  strategy, real marketing um is is where I would [38:33] use that having those real conversations, finding  out how we can make an impact in the market. It's [38:38] exactly kind of what close I was going to say,  the conversations, talking to people. I have 150 [38:42] members across the country that I don't get  to talk to enough that I want to learn from. [38:46] They're on the ground. They're doing the work.  And that insight is what helps me build better [38:51] strategies at the national level. And I don't have  as much time as I would love to do that today. So, [38:55] if I had 10 hours, it's me talking to more CEOs  and local markets, understanding more about their [39:00] challenges and their issues and how we solve them.  Uh, I was thinking about hot baths. People are [39:08] so overwhelmed right now. People are seriously so  overwhelmed right now with having to stay ahead to [39:15] learn. I mean, it's it's a full-time job to spend  time learning all of this on top of your job and [39:21] then driving your own personal change management  to get there. I legitimately think people need [39:26] more space. Uh just empty space to explore or or  for example back in the day, right? There used [39:32] to be time for programs um where you could have  people do their own inspirational projects. Let's [39:38] get back to that. Let's let's use those hours to  get back to things like that and and great ideas [39:42] will come out. I'm going to split my time 50/50  of it. I am such an AI nerd that I'm going to be [39:50] sitting and playing with all the tools because  there's new ones every day. and I feel like I [39:54] just want more time to play. Um, so half my time I  would do that and then the other half I would just [40:00] I would spend it talking to customers and buyers  and building relationships and understanding their [40:06] day-to-day problems and what keeps them up at  night. I think that you can never do enough of [40:10] that. All right, I think that's good inspiration  for all of us. Um, we have about three minutes [40:16] for questions. Couple questions right here. And  I think we've got a microphone right behind you. [40:24] All right, here we go. You spoke earlier about the  need for data quality and completeness. What sort [40:31] of AI are you leveraging to manage this process to  fill in gaps without sacrificing quality? Does who [40:40] has a solution for that? I seriously I I think  you're right. We're still trying to figure that [40:46] out. We we have taxonomy, layers, skimming, all  of it. And and right now it's people still trying [40:55] to figure that out, but we've got to get to the  automated piece or we're going to spend years and [40:59] years and years trying to get the data right  and then finally be able to use it. So we we [41:04] have to expedite and scale that significantly.  I can speak to it just for my use case, which [41:09] is local market intelligence, right? So inside  businesses, you might have a little bit of data. [41:15] So being able to leverage all of that signal data  and do it at a statistical significant level, [41:21] that's what we typically do and we we are able to  smooth out some of those signals. Um but again, [41:26] you're going to find a lot of platforms depending  on what kind of data you're talking about there [41:30] there are data pipes that can help with  that. Well, and if the quality is not there, [41:34] you're just putting together Yeah. big stuff.  Time for one more question. This gent over here [41:45] Thank you. Um, really good followup actually.  So, I'm Matt from Poff Labs. We help companies [41:51] aggregate their data together. Um, so they  can build their own customer brains. So, [41:55] my c question is for Kathleen. Um, for a company  brain, this is actually multi-layered. The first [42:01] one is what happens when your data sources  disagree with each other? say your CRM notes [42:06] disagree with um your account or your CDP and  how do you handle aggregating your fragmented and [42:13] unstructured customer feedback because customer  feedback is never onetoone verbatim to surface [42:18] insights across um the platform reliable rei  reliably. Yeah, it's a great question. So as [42:25] I was describing we the interface for this this  solution we built is in Slack. The responses are [42:31] very visible and we have conversations with it.  It's actually kind of funny because we built it [42:35] and launched it and it it has a personality.  It's kind of crazy. Um so we we can really [42:42] interact question. It will tell us when it's not  sure. It will give us the percent probability of [42:46] confidence on its response. Um and our team can  jump in and correct it. Um but the other thing I [42:52] would say in response to your question is there  is nuance in in how you access your data. And [42:58] um so I I publish a Substack where I write about  how I'm working with AI. It's called Code Meets [43:03] Creed. I literally just published an issue this  morning about for me what a game changer vector [43:08] search has been. So I've been putting a lot of  my raw data into Supabase and I just layered on [43:15] vector search to that. And for anybody who doesn't  know what I'm talking about, I am not an expert in [43:19] this but the way I describe it is like a regular  database. You can search for exact word matches. [43:25] Vector search lets you search for meaning and it  it it is more intelligent. And that has been a [43:30] game changer for me to be able to find nuance and  context within the data. I would also just add you [43:36] have to know who your brand is, right? So if those  things don't line up and you're absolutely right, [43:41] what people say is gold, but at the end of the  day, you're still trying to tell your story. So [43:45] it should always use or reference back to that  brand brain to say what do we actually want to [43:51] say and how do we prioritize this? And that can  be connected quite easily as well. All right, [43:57] we are over time. So, we are done. Thank you  everyone. Thank you all so much for coming.