1mind AI Agents Scale Go-to-Market Funnels to Drive 16x ROI

1mind AI智能体规模化扩展GTM漏斗，实现16倍ROI 以自主AI智能体取代传统多阶段销售交接，实现买家即时推进，并将单次咨询的收益产出提升至三倍。 Jonathan Kvarfordt深入探讨了从以销售为中心的销售工作流向以买家为中心的AI执行模式的转变，直面传统多阶段交接中固有的结构性瓶颈。他重点阐述了1mind如何运用自主运行的“Superhumans”，在包括网站访问、实时通话和支持队列在内的整个漏斗流程中与买家互动。通过提供即时解答、无门槛探索和实时互动演示，GTM团队相较于传统销售模式能够显著提高转化率和ROI。 1mind目前在70多个品牌的整个买家旅程中，已有超过100个自主运行的“Superhumans”投入实际生产运营。 (0:04) 部署1mind智能体AI技术的GTM团队在极短时间内实现了高达16倍的ROI。 (0:04) 演讲者Jonathan Kvarfordt（“Coach”）的工作重心已从在Momentum优化销售人员工作流，转向在1mind优化买家体验。 (0:04) 由于获客成本急剧上升且收益递减，依靠增加销售人员数量来扩大营收的传统GTM模式已难以为继。 (2:05) 最新研究显示，94%的买家在直接与企业接洽之前会利用AI进行独立调研。 (2:05)

Jonathan Kvarfordt, VP of Marketing at 1mind, argues that traditional go-to-market models are failing because they prioritize internal seller workflows over buyer convenience, necessitating a transition toward autonomous AI agents that engage prospects immediately. Refocusing GTM Strategy on Buyer Experience (0:04) Drawing on his background leading marketing and strategy at Momentum—a sales enablement platform acquired by Salesforce—Kvarfordt contrasts legacy GTM software with emerging agentic frameworks. Traditional tools focus almost exclusively on internal seller operations, such as automating CRM updates, call recording, and rep coaching. In contrast, 1mind targets buyer-facing friction by deploying autonomous AI agents—termed "Superhumans"—designed to provide instant answers, ungated product exploration, and self-directed progression. Structural Flaws in Headcount-Driven Sales Funnels (2:05) Citing research from Benchside, Kvarfordt observes that the historical playbook of scaling revenue by expanding sales headcount has become economically unsustainable. Adding personnel has driven customer acquisition costs upward while diminishing returns, prompting industry-wide restructuring and headcount cuts across revenue organizations. This structural strain is compounded by evolving buyer behaviors: an estimated 94% of buyers now utilize AI for independent research prior to initiating contact with a vendor. While these buyers expect transparent information and autonomy, traditional enterprise models still force them through a sequential chain of human handoffs—from SDRs and BDRs to AEs, sales engineers, implementation specialists, and customer success managers—a division of labor originally established solely to accommodate human cognitive and logistical limits. Eliminating Scheduling Friction Through Instant Progression (16:12) Kvarfordt identifies the standard requirement to schedule future discovery meetings as a major source of funnel drop-off. By allowing an automated system to immediately ask buyers if they want an instant demo and launching directly into substantive dialogue upon consent, organizations can capture buyers at peak interest. Kvarfordt claims that 1mind's live deployment of more than 100 autonomous agents across over 70 brands has yielded returns of up to 16x ROI by eliminating human follow-up delays. While Kvarfordt highlights rapid ROI and immediate conversion gains from replacing rigid human handoffs with autonomous agents, the broader model relies on whether AI-led interactions can consistently satisfy complex enterprise evaluation criteria across the entire customer lifecycle.

1mind 市场营销副总裁 Jonathan Kvarfordt 认为，传统依赖人力线性分工的销售模式已无法适应当下买方自主调研的趋势，企业应通过部署全链路 AI Agent 将重心从优化销售人员操作转向满足买方的即时体验。 从优化销售人员转向重构买方体验 (0:04) Jonathan Kvarfordt 指出，传统市场拓展（GTM）策略过度聚焦于销售端人员的内部效率优化，例如 CRM 更新、电话沟通记录与销售辅导，却忽视了实际采购者的摩擦成本。他对比了以往在 Momentum（后被 Salesforce 收购）与 1mind 的实践，提出应转向以买方为中心的新范式，为客户提供无门槛的产品探索、即时问答以及自主推进采购流程的通道。据其陈述，1mind 目前有超过 100 名自主 AI 智能体在 70 多个品牌处于实盘运行状态，并称部分团队借此在短时间内获得了高达 16 倍的投资回报。 传统线性销售漏斗的人力与成本瓶颈 (2:05) Kvarfordt 引用 Benchside 的研究指出，依赖堆砌人手来拉动营收增长的旧模式已陷入瓶颈，客户获客成本（CAC）攀升而边际产出下降，直接导致了近期行业内的团队规模调整。与此同时，采购方行为已发生结构性转变：调研显示 94% 的买方在与企业接触前会使用 AI 进行独立调研，期望获得透明、直接且无冗余信息的交互体验。然而，传统销售体系仍强行使用 SDR、BDR、AE、SE 等层层转交的流水线模式；这种模式本质上是过去为了应对单个人力无法跨周期维护成千上万客户上下文而设立的权宜之计。 即时交互与演示消除流失并提升转化 (16:12) 针对传统销售链路中等待会议排期造成的客户流失，Kvarfordt 强调在买方意向最高点提供即时满足的价值。当系统向访客提供即时启动产品演示的选项并获得许可后，AI 能立即开展具有实质深度的交互与解答。这种机制消除了等待人工销售代表跟进的时间差与阻力，从而在采购需求最迫切的时刻锁定商机，推动回报率的快速释放。 本场演讲主要阐述了转向以买方为中心的 Agentic AI 运行逻辑，但现场因音视频播放受限而未能完整展示具体产品工作流机制，其大规模落地的稳定性和更深层的数据边界仍有待进一步实际验证。

## Background 本内容由 Ai4 于 2026 年 8 月 24 日在 YouTube 平台发布，演讲者为 1mind 市场营销副总裁 Jonathan Kvarfordt（Coach）。页面简介指出，本次演讲深入剖析了企业级 Agentic AI 跨整个买家旅程规模化落地的运营模型，展示了 100 多个 Superhuman 智能体如何依托共享上下文图谱与治理护栏，在 CRM、产品及营收系统中协同作业，并为客户带来高达 14 至 16 倍的 ROI。 ## TL;DR Jonathan Kvarfordt 提出 B2B 营销正从“武装销售”转向“赋能买家”，直击传统线性销售流程中的人类生理与认知瓶颈。他深入解析了构建企业级 Superhuman 智能体时在准确度、功能广度与低延迟之间的权衡三角，并展示了智能体如何接管深度问答、即时 Demo 演示和随行技术支持（Ride-Along SE）。通过 HubSpot、Alteryx、Seismic 及 ZoomInfo 等品牌的实盘案例，证明全流程 Agent 协同不仅能够实现成倍的“产能飞跃”，更能将人类销售释放至高情商的信任构建环节。 ## Atmosphere & style 整场演讲洋溢着一线操盘手（Operator）务实、极客且充满激情的实战气息。Jonathan 语言风趣犀利，以“Demo 猴子”等生动比喻直击行业痛点，毫不避讳传统虚荣指标与技术权衡的残酷现实。现场节奏紧凑、案例翔实，将前沿理论与扎实的业务数据无缝熔接，给听众带来极具说服力与启发性的未来工作流冲击感。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=4s) 1mind 在生产环境中部署了 100 多个 Superhuman 数字智能体，助力企业客户在数月内实现了高达 16 倍的投资回报率（ROI）。 - [01:30](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=90s) B2B 采购范式发生剧变，高达 94% 的买家已通过 AI 自主完成前置调研，传统依赖人力层层设卡（Gatekeeping）的销售模式陷入了成本倒挂与增长瓶颈。 - [03:23](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=203s) 传统分工流水线（SDR/AE/SE/CSM）本质上是对人类生理与认知极限的妥协，仅靠 Copilot 进行局部修补无法带来质变，必须自顶向下重塑买卖双向流程。 - [07:24](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=444s) 构建高阶企业级智能体必须平衡由准确度（Accuracy）、功能广度（Breadth）与响应延迟（Latency）构成的不可能三角难题。 - [09:40](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=580s) 摆脱单纯追求“预约会议”的虚荣指标，让智能体在前链路直接消化深度答疑、动态 Demo 与资质审查，能够有效过滤无效线索并交付更高价值的成熟 Pipeline。 - [12:00](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=720s) 商业成交的核心驱动力是信任，1mind 推出的 AI 随行技术顾问（Ride-Along SE）能在会议中实时解答技术难题，打破了 SE 人力配比不足的规模化瓶颈。 - [14:44](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=884s) Agentic AI 的核心价值在于实现“产能飞跃（Capacity Lift）”，通过覆盖单位经济效益不足的边缘盲区实现业务扩容（例如在 HubSpot 单个智能体完成相当于 83 名 SDR 和 19 名 SE 的工作量）。 - [16:12](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=972s) 向进站买家提供即时专属 Demo 能够即时满足客户诉求并显著缩短交易周期，助力 Seismic 等企业在上线首月获得超百万美元的有效商机。 - [16:12](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=972s) Alteryx 落地了由 Annie（官网互动）、Atlas（实时销售技术支持）和 Adam（客户成功与入职）组成的全流程 Superhuman 协同智能体矩阵。 - [17:02](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=1022s) 全流程 Agentic AI 的核心哲学在于人机协同：前端由 AI 实现零延迟的即时信息交互与流程推进，后端释放人类员工专注于高情商（EQ）的深度信任与伙伴关系建设。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=4s) 个人背景与 1mind 的核心使命 Jonathan Kvarfordt（业内人称“Coach”）以 1mind 在市场上已实盘运行的 100 多个 Superhuman 数字智能体切入，分享企业级 Agentic AI 如何在短时间内为客户实现高达 16 倍的投资回报率（ROI）。谈及自己的绰号来源，他分享了一段早年在初创公司的趣事：当时 40 个人挤在一间狭小的办公室里，竟然有 20 个人的名字都叫 John，为了区分大家便纷纷起了外号，由于他负责销售辅导，"Coach" 这个称呼便一直沿用至今。 在职业履历上，Jonathan 拥有深厚的 RevOps（营收运营）与销售赋能背景。五年前他创办了拥有超过一万名学员的 GTMAI Academy，曾在近期被 Salesforce 收购的 Momentum 担任市场与策略副总裁，并在布莱恩大学（Brian University）讲授面向职场人士的 AI 专业课程。他坦言自己并非无所不知的学术权威，而是一个彻头彻尾的“AI 极客”和一线实战操盘手（Operator）。在 Momentum 的两年间，他的核心职责是优化销售人员与客户成功团队的工作流；而加入 1mind 后，他将视角彻底调转——全力探索如何优化“买家”的采购体验。 --- ### [01:30](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=90s) 范式转移：从“武装销售”转向“赋能买家” 当前的 B2B 科技领域存在一个巨大脱节：绝大多数工具都在竭力武装销售人员，买家的真实体验却被抛诸脑后。正如行业研报所指出的，现有的商业增长与成本模型正在陷入倒挂困境：依靠不断堆砌人力和扩张成本来换取增长的旧模式难以为继，导致成本不断攀升而增速持续放缓，促使各大企业纷纷进行成本结构调整。 与此形成鲜明对比的是买家行为的剧烈变化。调研数据显示，高达 94% 的买家已经在利用 AI 工具自主完成前置研究，对市场选项了如指掌。当他们真正访问企业官网时，需要的不再是销售人员的推诿与层层把关（gatekeeping），而是即时、精准、有深度且充满同理心的解答与掌控感。 ```mermaid flowchart TD accTitle: Traditional vs Buyer-Centric AI Model accDescr: Comparison between legacy seller-centric linear pipeline and modern buyer-centric autonomous model. subgraph Legacy["传统线性模式 (受限于人力瓶颈)"] direction LR A["SDR / BDR 筛选"] --> B["AE 演示建联"] B --> C["SE 方案解答"] C --> D["CSM / AM 交付续约"] end subgraph Modern["1mind 赋能买家模式 (Superhuman 驱动)"] direction TB E["自主买家进站 (94% 已完成前置研究)"] --> F["Superhuman 全能交互 (即时 Demo / 深度问答 / 自动资格判定)"] F --> G["高意向成熟商机无缝切入人工销售"] end ``` --- ### [03:23](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=203s) 传统销售流程的死穴：人类产能瓶颈与系统重构 经典的 B2B 销售流水线（SDR 寻找线索 → BDR 确认意向 → AE 推进交易 → SE 方案支持 → 实施交付 → CSM/AM 运营续约）之所以被切得如此细碎，并非因为这是最佳体验，而是为了妥协人类个体的生理与认知极限：单个员工无法在面对成千上万客户时全盘兼顾。 人类本身成为了流程中的核心瓶颈。这并不是要全盘替代人类员工，而是必须认清人类在规模化扩张中的局限：人力成本高昂、脑容量有限，而且正如销售赋能管理者所熟知的，“销售代表在沟通中经常会产生幻觉（sales reps hallucinate all the time）”。当前企业在拥抱 AI 时分化为两条路径：一种是抱残守缺，在陈旧的销售流程上强行贴上 AI 标签，仅用 ChatGPT 或 Copilot 帮员工更快地写写邮件，这种自下而上的局部修补无法带来质变；另一种则是基于“待办任务”（Jobs-to-be-Done）框架，由管理层自顶向下推动系统级重构，同时重塑买卖双向流程。 --- ### [05:12](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=312s) 交通革命类比：自顶向下的买家体验重塑 正如现代出行方式的多样化发展——人们既可以租车、骑车，也可以打 Uber 或乘坐 Waymo 等自动驾驶汽车一样，企业服务也迎来了自主驱动（Autonomous）的新纪元。真正的数字化变革必须由领导层自顶向下推动，在保障买家体验的前提下重新分配人机协作的权重： - **局部提效的局限**：单纯为销售配备 Copilot 只能加快写邮件的速度，无法引发根本性的营收模式创新。 - **系统重塑的契机**：让 AI 承担全天候的数据吞吐与基础交互，在关键信任节点引入人工专家，打造兼顾自主与温度的采购路径。 --- ### [07:24](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=444s) 突破“预约会议”怪圈与企业级 AI 的“不可能三角” 市面上绝大多数网站聊天插件的底层逻辑依旧局限在“诱导约会”（Book a meeting），本质上只是在把买家往人工流程里推。1mind 提出的核心命题是：如果让 Superhuman 智能体承担更多职责，在 7 步销售流程中直接消化前 3 步——即时完成深度问答、动态 Demo 演示以及资质审核，让买家在接触人工销售时就已经处于决策深水区，整个交易周期将被大幅压缩。 在实际案例中，Alteryx 的首席营收官（CRO）曾先后 4 至 5 次在 1mind 官网上与 Superhuman 深入交流，其中一次单席对话长达 90 分钟。当他发起正式采购申请时，1mind 的 CEO 第一时间介入跟进，对方开门见山地明确了所有部署场景，使得原本漫长的企业级大单在短短几个月内迅速签约落地。 构建此类高阶企业级智能体必须克服由三者相互博弈构成的技术难题： | 核心维度 | 挑战与技术博弈 | |:--- |:--- | | **准确度 (Accuracy)** | 追求极致响应速度时，模型输出的精准度极易下滑，产生业务幻觉。 | | **功能广度 (Breadth)** | 赋予智能体调取跨系统后台、动态演示与深层逻辑运算时，会显著增加系统延迟。 | | **响应延迟 (Latency)** | B2B 线上买家的等待耐心极其苛刻，响应若超过 10 秒买家便会直接流失离场。 | --- ### [09:40](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=580s) 真实对抗：从虚荣指标到高质量管线与 PLG 破局 在一组 50/50 流量的 A/B 对抗测试中，传统竞品 Chatbot 虽然依靠强行索取信息录得了更多的“会议预约量”，但 1mind 的 Superhuman 在后端激活的真实有效 Pipeline 价值却呈现出压倒性优势。Superhuman 在前面对不契合的线索进行了自动分流与淘汰，从而避免了销售团队把精力白白浪费在毫无转化前景的无效会议上。 在产品驱动增长（PLG）场景中，以 HubSpot 为例，其大量中小微客户（SMB）过去受限于人力成本而无法获得专属销售或支持服务。部署 Superhuman 后，AI 能够在单次连贯交互中引导用户完成“免费试用注册 → 购买签约 → 系统入职配置（Onboarding）”的全生命周期闭环。现场通过 Alteryx 数字员工 Annie 的演示展示了这一能力：AI 能够根据用户提出的银行风控场景，实时调用后台模块完成沉浸式走查与操作演示。 --- ### [12:00](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=720s) 信任驱动成交：AI 随行技术顾问（Ride-Along SE） 传统销售观念强调“人类只向人类购买”，但现实商业中更精准的法则是：“人类在产生信任时才会买单（Humans buy when they trust）”。无论面对大额还是小额采购，只要品牌、安全和解决方案建立了足够的确定性，客户就愿意推进决策。 为了在技术销售环节构建深度信任，1mind 推出了可直接接入 Zoom 会议的“AI 随行解决方案架构师（Ride-Along Sales Engineer）”。在实际访谈中，人类技术顾问（SE）不仅没有感到职业威胁，反而给予了极高评价： > “我再也不用当个‘演示猴子’（demo monkey）去一遍遍回答基础机械的问题、重复那些千篇一律的演示了。我终于能把精力集中在梦寐以求的创造性、战略级解决方案设计上。” 企业过去受制于高昂成本，极难实现 SE 与 AE（客户经理）的 1:1 满额配比，而 AI 随行架构彻底解决了技术演示能力的规模化瓶颈。 --- ### [14:44](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=884s) 产能飞跃（Capacity Lift）：HubSpot 与 ZoomInfo 的落地实证 企业落地 Agentic AI 的最佳切入点并非盲目替代既有人力，而是瞄准那些**因人力单位经济效益受限而覆盖不足的高意向、低转化盲区**。通过这种方式实现的业务扩容被称为“产能飞跃（Capacity Lift）”。 - **HubSpot**：单个 Superhuman 智能体完成了相当于 83 名 SDR 和 19 名销售工程师（SE）的工作量，期间没有裁撤任何一名员工，却达成了用户平均交互时长 7 分钟、整体营收增长 25%、免费试用转化率达 70% 的全流程贯通成果。 - **Seismic**：在其官网底部提供“观看录播视频”、“预约人工沟通”与“立即获取专属 Demo”三项路径，仅通过开放即时 Demo 这一选项，便在上线前 30 天内斩获了超过 100 万美元的有效商机管线。 - **Experity**：将 1mind 部署于对抗测试中，在极短时间内实现了赢单率与销售周期的成倍优化。 - **ZoomInfo**：在不改变原有获客架构的前提下，将 Superhuman 智能体嵌入到留资表单的核心触点后承接即时交互，在 5 个月内实现了高达 16 倍的投资回报率（16x ROI）。 ### [16:12](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=972s) 即时 Demo 转化与 Alteryx 的 Superhuman 落地矩阵 在客户旅程中，与其让高意向潜在客户填写表单并苦等几天后的销售会议，不如直接在当下向他们提问：“你现在想看演示吗？”当访客选择“是”时，AI Agent 便会立刻引导其进入深度的互动与产品对话中。这种为潜在客户提供即时满足、无需等待人工排期的体验，让企业在极短的时间内获得了巨大的投资回报率（ROI）。 以知名数据分析平台 Alteryx 为例，他们在业务全流程中部署了三到四个核心 Superhuman（智能体员工）组成的协同矩阵，各自承担不同的关键职能： ```mermaid flowchart LR accTitle: Alteryx 协同智能体工作流架构 subgraph S1["网站交互与转化"] Annie["Annie<br/>官网互动与即时答疑"] end subgraph S2["实时销售支持"] Atlas["Atlas<br/>实时同席销售助手<br/>(解答复杂技术问题)"] end subgraph S3["客户成功与入职"] Adam["Adam<br/>CSM 与新客户 Onboarding"] end Annie --> Atlas Atlas --> Adam ``` - **Annie（官网智能体）**：驻扎在 Alteryx 官方网站，负责第一时间接待访客，展开深度对话并驱动初步转化。 - **Atlas（实时同席助手）**：作为“实时跟跑（Ride-along）”智能体加入线上销售会议，在销售代表与客户通话过程中实时提供技术支持，即时解答深度的技术难题。 - **Adam（客户成功智能体）**：专注于客户成功（CSM）与新客户入职引导（Onboarding），保障成交后的交付与持续使用体验。 除了 Alteryx，演讲者还展示了多份不同规模合作企业的匿名化部署数据，以证明该方案不仅适用于单一场景，更能够适配不同体量企业的全流程业务运转。 --- ### [17:02](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=1022s) 落地验证与人机协同愿景：兼顾即时解答与高情商联结 面对行业中“到底有谁真的需要全流程 Agent 协同”的质疑，演讲者强调，这绝非停留在 PPT 上的未来理论或假设，而是已经在生产环境中稳定运行并产生极佳业务回报的成熟模式。全流程 Agentic AI 的终极目标不是完全取代人类，而是对买家体验进行彻底重塑，实现两全其美的人机协同模式。 一方面，在客户产生疑问的当下，AI 能够第一时间提供所需的精准答案，消除流程中的等待摩擦，大幅加速购买进程；另一方面，它把宝贵的人力释放出来，让业务人员将精力集中在最需要人类发挥“高情商（EQ）”与温度的环节去建立真实信任与深度伙伴关系。 - **生产级验证**：多 Agent 协同系统已在数十家品牌落地，彻底打破了仅能处理单一用例的局限。 - **核心哲学**：前端以 AI 提供零延迟的即时信息输入，后端由人类专家主导高情商的关系维系。 - **现场互动与展台指引**：演讲者在结尾邀请现场对 Agentic AI 落地感兴趣的 GTM 与技术负责人前往 **853 号展台**进一步交流并领取晚间活动门票，随后进入现场 Q&A 环节。 ## Notable quotes > [03:23](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=203s) 销售代表在沟通中经常会产生幻觉（sales reps hallucinate all the time） — Jonathan Kvarfordt > [12:00](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=720s) 人类在产生信任时才会买单（Humans buy when they trust） — Jonathan Kvarfordt > [12:00](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=720s) 我再也不用当个‘演示猴子’（demo monkey）去一遍遍回答基础机械的问题、重复那些千篇一律的演示了。我终于能把精力集中在梦寐以求的创造性、战略级解决方案设计上。 — 人类技术顾问（SE） > [16:12](https://www.youtube.com/watch?v=H2Z-jUnGyqg&t=972s) 你现在想看演示吗？ — Jonathan Kvarfordt ## Mentioned Jonathan Kvarfordt (Coach), 1mind, Superhuman, Ai4, GTMAI Academy, Momentum, Salesforce, Brian University, HubSpot, Alteryx, Annie, Atlas, Adam, ZoomInfo, Seismic, Experity, Pipedrive, Uber, Waymo, Zoom, ChatGPT, Copilot ## Video description Jonathan Kvarfordt - VP, Marketing, 1mind Standing up one agent is easy. Running a coordinated workforce of them across your entire buyer journey, on guardrails, in production, is the part almost nobody has solved. This session takes GTM and AI leaders inside a real deployment: 100 Superhumans live across 70+ brands, working the website visit, the live call, the free trial, the support queue, and the renewal from one shared brain. You will see how the orchestration layer, the account-level GTM Context Graph, and human-in-the-loop guardrails hold up when Superhumans read, act, and write across CRM, product, and revenue systems at once. Then you will see the results that follow: one go-to-market data company hit 14x ROI in three months, Experity drove 3x more revenue per 100 inquiries, and Pipedrive doubled its trial conversions. What you will leave with: 1. The operating model for scaling past a single use case, where one shared context graph makes every Superhuman smarter as you add the next. 2. How to govern autonomous systems acting across CRM, product, and support, using guardrails, source citations, and human-in-the-loop patterns that survive production. 3. What actually breaks at scale, including latency, cost, reliability, and drift, and the trade-offs teams make to keep every buyer experience instant. 4. How the funnel scoreboard changes at scale, and the revenue, ACV, and cycle-time metrics to watch instead of meetings booked. [Source](https://www.youtube.com/watch?v=H2Z-jUnGyqg)

[3:23] Jonathan Kvarfordt: 销售代表在沟通中经常会产生幻觉（sales reps hallucinate all the time） [12:00] Jonathan Kvarfordt: 人类在产生信任时才会买单（Humans buy when they trust） [12:00] 人类技术顾问（SE）: 我再也不用当个‘演示猴子’（demo monkey）去一遍遍回答基础机械的问题、重复那些千篇一律的演示了。我终于能把精力集中在梦寐以求的创造性、战略级解决方案设计上。 [16:12] Jonathan Kvarfordt: 你现在想看演示吗？

Jonathan Kvarfordt: Standing up one agent is easy. Running a coordinated workforce of them across your entire buyer journey, on guardrails, in production, is the part almost nobody has solved. Jonathan Kvarfordt: 100 Superhumans live across 70+ brands, working the website visit, the live call, the free trial, the support queue, and the renewal from one shared brain. [16:12] Jonathan Kvarfordt: Do you want to have a demo now?

[0:04] Today, if you're not here for this, hopefully the  what we're going through is the 100 superhumans [0:07] live. I'm going to go through some use cases of  one mind, what we've done across what we have 100 [0:12] superhumans live in the market right now. And  I want to show you how teams have hit uh 16x [0:17] ROI in very quick time using some of the tech we  have. So, and hopefully I have some videos. They [0:22] may or may not work. I may have to just tell  you versus show you sadly, but we'll we'll go [0:26] through it and my team if you're interested  can show you some things at the booth later. [0:30] All right, so real quick about me. My name is  Jonathan Carford. Some people call me coach. [0:34] Long story short, I was in a startup. There  was 40 of us in a very small room. 20 of us, [0:38] no joke, were named John. And so we all got  nicknames. Mine was coach because I was a [0:42] sales coach to the team and then from there it  just stuck. And so people just called me coach. [0:46] I'm currently a VB market one mind. I'm the  found founder of the GTMAI Academy 5 years ago [0:52] and um I was the VP of marketing and strategy at  Momentum who was recently acquired at Salesforce. [0:58] Had the option to go to Salesforce, great team.  I just love the startup world. So I decided to [1:02] go back to the startup life and I've worked  with a ton of other startups. I also teach at [1:06] Brian University of AI for professionals  and there are professional courses. So [1:10] I live, breathe, eat, teach, sleep, talk AI  all the time. I'm definitely not an expert. [1:16] I am definitely a nerd. So hopefully you can  understand from my operator point of view that [1:21] I'm very passionate about this. I'm trying to  learn all the time what I can do better and being [1:25] places like this and listening to brilliant people  is a lot of fun for me. So really glad to be here. [1:30] Um with Momentum to kind of set the the frame  of what we're going to be going through. [1:36] My job on Momentum was to optimize the sellers.  All we did was look at what's the seller's [1:40] workflow or the customer success workflow and how  do we make that as easy and as fluid as possible. [1:45] So all I did for two years and my background  is in revops and enablement was focus on that. [1:50] With one mind it's shifting gears now to where  I'm looking at how do I optimize the buyers. [1:55] Um and this is where I think we have a disconnect  in a lot of the technologies out there is that [1:59] people focus a lot on how do I optimize the  seller and the buyer gets kind of lost in the mix. [2:05] So from bench side they put out some reports  that talked about how we're in this place kind of [2:10] laying the land where growth and costs are going  up upside down to where in order to grow more [2:17] the old model was to hire more or to expend more  cost which means that the you obviously your cost [2:22] goes up and your growth goes down. This obviously  cannot be sustainable which is why you see I think [2:28] all these corrections in a lot of teams trying to  figure out where they have too much cost or not. [2:34] Um, from there you have like on talking about  buyers. If you think about your own buyer's [2:38] journey when you go through conferences like  this or going online or whatever the case is, [2:43] more than likely you probably feel the same  way we found what people want more control [2:46] over what they're finding or not. They want to be  able to have answers. Now, we've been trained for [2:51] put out research that talked about how 94% which  literally is everyone uses AI to do research which [2:57] means that we are much more intelligent and aware  of what options we want to talk to. So by the time [3:03] we actually get to a website or get to someone,  we're looking for more than what we're getting. [3:08] People want more accuracy instead of just fluff.  And then um couple more things is no gatekeeping. [3:14] I don't think you see the lower one. Um but they  want more gatekeeping and more empathy for what [3:18] they're going through and people who understand  their pain of what we experience as buyers. [3:23] There it is. So here's a sales process.  Anyone's look familiar to anyone here? Anyone [3:29] have something like this? So you have uh SDR, BDR,  AE, SE, solution engineer, implementation, CSM, AM [3:35] and you have these stages that we go through.  My question is why do we do it this way? [3:41] And the answer is because this was what we had  to do because we had human limits. We had to [3:47] operationalize around because every person like  one person couldn't handle all of this at one [3:51] time when you're scaling up for thousands of  customers. So we did this because we had to. [3:57] So humans alone are the bottleneck. I'm not  saying we should fire humans. That's not [4:02] my argument. My argument is is if you  look at what is possible now with AI, [4:07] there's a lot more options to grow exponentially  and use humans where our superpowers are, where [4:11] we need them most and use AI where AI can now do  things we've never had been able to do before. [4:16] So the the side with humans is like  if you think about from a scaling, [4:20] we are a little bit more expensive. Humans  hallucinate all the time. Like I was in [4:24] enablement and coaching. I can tell you sales  reps hallucinate all the time. So very very common [4:29] experience and we have limited capacity. Like I  can only handle so much in my brain at one time. [4:35] So we want to look at like the two realities that  I see when I talk to people in the academy. I've [4:39] had 10,000 people plus go through the academy.  I do strategy sessions with CRO all the time [4:44] and I see two models happening. One of which  is they want to do the old way with AI. [4:49] Old sales process, old buyers process. just slap  AI on it and do it faster, better, cheaper, etc. [4:55] And there's other teams who looking at how do  I do this in a totally different way? How do [4:59] I do this from a system point of view or  um jobs to be done framework so that you [5:04] rethink both how the seller works and then how  the buyer works for the best of both worlds. [5:12] So okay, so the strategy is when I do strategy  sessions and in fact yesterday I did a a [5:16] presentation on AI strategy that kind of talks  about this there's two versions. One is obviously [5:21] using the the ground up. Um what I find is when  a lot of teams throw co-pilot or chatbt at a a [5:26] sales team member or CS or whoever it is they're  working, um you can only get so much productivity [5:32] from writing an email faster. It doesn't create  true transformation. In order to have true [5:36] transformation, you need to have leaders from the  top down to be able to be open for something new [5:41] and to um attempt to do something for the  benefit of the buyer that will require change. [5:48] I think you need to have the best of both  worlds. I think people, individuals need [5:51] AI at their disposal, but again, if we want  massive change, it has to come from leadership. [5:57] Okay. Sorry, I don't this thing is working.  There we go. Um the the system, if you think [6:02] about the seller, and I can say this because I  was at Momentum and I sold this for two years. [6:06] Um is all about the seller. How do I optimize  the seller with recordings, coaching, [6:11] social engagement, CRM, maps, technology, all  this stuff? And the other side is but what about [6:16] what about the the um buyer? How can we get agents  or superhumans or whatever AI to handle the things [6:22] they need to have and then bring in humans  at the most important critical part for the [6:26] relationship to make sure it can move forward.  Uh we like to compare this to Uber and toe. [6:32] You have when you travel you have options.  You can either rent a car, get a bike, [6:36] have a friend get you, get an Uber, get a  Whimo, depending on the city, and you have [6:41] options as as a buyer of transportation. We now  have options of autonomous drivers, and we have [6:46] the Tesla car out here showing their new stuff to  be launching here soon. Um, we have options and [6:52] we can figure out where do we want to go, which  is changing the entire transportation industry. [6:58] So, come on, work with me. There we go. I of all  the companies I've worked with with thousands [7:03] through both one mind and momentum and through  the academy most teams are not built for the [7:07] AI world and mostly because we've never done  this before. We're all pioneering together. [7:12] So I want to talk about the new experience and  the new way of of how you can think about this. [7:17] Um and even if you use our tech or not hopefully  I can give you some insights of what you could do [7:21] hopefully when you leave here afterwards. [7:24] So think about the process from a buyer's point  of view. And if they meet a chatbot of some kind, [7:30] whether it's advanced or not, whether it  has a face or not, the goal of that agent [7:34] is to book a meeting. It's operationalizing and  trying to optimize someone to talk to a human, [7:39] which again, there's nothing wrong with that, but  they're trying to looking at the customer journey. [7:43] They're trying to make sure that that keeps  going. The question we ask in one mind is what if [7:49] you could have a superhuman or an AI that goes  through and does more than just qualify, books a [7:54] meeting, and then hands it off to an STR and then  goes through the same rigomeroll we did before. [7:58] What if it could do more? What if it could  do demos and answer questions and qualify [8:03] and go down the path? Like if you have a sevenst  step sales process and if AI could do the first [8:08] three steps of those so that by the time a  human gets to another salesperson they're way [8:13] farther down the process and they're much faster  through the process than they would be before. [8:18] example from us from our own technology is  um the CRO of Altrix came in and talked to [8:23] our superhuman on the website four or five times.  one of which was 90 minutes long. And by the time [8:28] he came inbound, my CEO saw he was coming. He's  like, I want to go talk to him. And he he talked [8:33] the entire time saying, "I want to use them here.  I want to use them here. I want to use them here." [8:37] And eventually came uh a customer within a couple  of months very quickly for a large deal. Like [8:42] enterprise deals don't usually work as fast  as they do with us because of people using [8:46] the technology to do part of the buying process  with AI. And we see this happen all the time. [8:53] Okay. So there's three things I think no matter  what you work with, you need to have in order to [8:57] make this work correctly for an enterprise type  agent. Um, and these all kind of fight together, [9:02] which is where we kind of focus with one mind.  But there's three main things. There's accuracy, [9:07] there's breadth, and there's latency. When you  have latency and you're trying to get speedy, [9:12] accuracy usually comes down. When you try to  have more capabilities, that creates latency. So [9:18] these three things are not easy to come together,  but they're needed for a conversation to happen [9:23] so that when a buyer comes to your website,  they're not waiting for 10 seconds because that [9:26] happens, they're gone. They're out of there. So  that's what we specialize on is trying to figure [9:31] out how do we get all three of these together in  a way that like moves things really forward in a [9:35] quick way so it's a good experience for the buyer.  There we go. Okay. So what we talk about is how do [9:41] we use superhumans and and figure out the friction  points in the process to reduce the friction for [9:46] the buyer and then bring in the humans at the best  part of the process for the benefit of the buyer. [9:52] For some companies who are PLG like for example  with HubSpot they had a segment of their um [9:58] uh buying process with people in SMB that they'd  never had support before and they couldn't scale [10:03] and headcount for it but no one was getting any  type of service. So they put a superhuman there [10:07] instead and now a superhuman takes them from  free trial to sign up to onboarding them in [10:11] one conversation which they never had before. So  depending on your enterprise motion or SMB it can [10:17] change but there's a lot of cool coverage you  can have. I like to compare it to to this like [10:22] you could theoretically get 11 labs an am tavis  there's tons of other technologies out there you [10:27] can do and throw a face on put on your website  and there you go. So an example is um this is [10:32] a competitor and this kind of proves the theory.  If a chatbot is focused on booking meetings, then [10:39] we went headto-head. So they gave us 50% traffic  on two different sites and we went headtohead [10:44] and um on our side you can see like they engaged  with ours more but the meetings we booked was less [10:50] even though the chat chatbot was way more. Now on  the back end though um the activated pipeline the [10:57] actual pipeline there's created was on our side.  So we booked less but they're worth way more. [11:01] And what we found was um the people who are  getting qualified out weren't a good fit or wasn't [11:07] the right time or they did the things and took  away that meeting from a sales rep who would have [11:11] just done the same thing anyways which is very  interesting because otherwise you have a lot of [11:15] meetings with sales reps that don't go anywhere.  It's a waste of time and energy and and resources. [11:20] Okay. So Oh good. It's working. Is it working?  Wait for it. Okay. I might have to I'm going to [11:26] talk about these things. So if you go to Altrix,  which you can do anytime, which this isn't going [11:29] to work for me. Um sound's not working, but this  just shows the demo of Annie. I'm might just talk [11:34] over so you can see it. This is someone asking  Annie for um the CR of Altrix. And as they're [11:39] talking, this is taking him through a banking um  use case and showing them a demo and showing how [11:44] to walk through the platform so they can see how  it works. Um I'm going to skip this since we're [11:50] just going to talk over them. So that's the  first one. The first use case is on the website, [11:54] which again is asking how far down the path can  we take someone if they're ready to go down. [11:59] The next one is um selling. So here's I'm  going to get you through real quickly. [12:04] My um you've probably heard the term more  the phrase around humans buy from humans. [12:09] And what we're seeing is I'm going to change  that a little bit. Humans buy when they trust. [12:14] How many things do you trust you've never talked  to a human before? Whether it's big purchase, [12:18] small purchase, whatever. How many times have  you bought something with no human in the loop [12:22] because you trusted trusted the brand, trusted  security, trusted whatever. So question is then [12:28] how do we create trust so that AI can actually  sell or at least help with the selling process. [12:34] Here's another one. This is  an example of live call of our [12:36] we have a um what's called a ride along that joins  zoom calls and actually does demos live. So in our [12:42] own team we have sales engineers and what we found  out and I interviewed my sales engineer last week [12:47] and I asked him like are you threatened by this  at all? He's like no I love it because I'm no [12:51] longer a demo monkey. I'm not going on and doing  the same questions, same demos I've always done. [12:56] I get to do the creative strategic  solutioning I've always wanted to do [13:00] and the ride along sales engineer like on  this call will take care of everything else [13:04] like the demos and and the advanced questions and  all that kind of aspects is really really powerful [13:09] which again the question is how many teams have a  onetoone coverage of sales engineers to ees it's [13:14] very few because it's expensive so the question  is how do we scale that and let the solution [13:19] engineers who are really good what they do focus  on what they're really good at versus just demos [13:25] um so this is my might work. Here we go. This  one's a live demo. I wish I could show this [13:28] because this is our thing we're launching  here soon, but imagine like a mixture of um [13:33] um Whisper Flow with the technology that  actually walk through a a uh back end. So, [13:38] if you could see this, which you can see at our  booth um in uh Notion, HubSpot, and some others, [13:44] this walks through the user's experience.  It says click here and either talk to them [13:49] or actually do it for them. So, the person  just talks and it'll fill out all the gaps [13:52] and do all the things for them, which is  really really wild. and pretty awesome. [13:58] Um, I'm going to skip through these. Sorry.  So, I want to give you some examples. Um, [14:01] Seismic is one of our customers and you can't see  on the bottom because of the the transcription, [14:06] but on the bottom of their website, they give  people three options. It's kind of like going [14:09] back to the transportation concept. What path do  you want to go down? And all they did was give [14:15] the human an option of getting a demo. Now,  they didn't take it away and focus on just [14:19] just having that experience. They said,  "Watch demo videos. um book a call with us [14:24] or get a demo now and they created out of  a million dollars in pipeline the first 30 [14:28] days by just giving the option of doing that a  more uh more farther experience. Experity had [14:34] another one where they put us up against another  AI chatbot where we doubled the results in the [14:38] first few days both the win rate sales cycle  etc. And then um let's keep going. There we go. [14:44] And the question people ask me is so where the  heck do I start with something like this? And [14:48] we like to first start by letting us prove it out  to people and saying where is a business model [14:53] where you can't have a human because of cost  time etc. So again HubSpot's a great example [15:00] the superhuman of HubSpot did a work of 83  SDRs and 19 sales engineers but those people [15:05] didn't exist there. We didn't fire anybody. No  one was moved. We just did the work of those [15:11] amount of people. I like to cover it uh call it  capacity lift where you can lift your capacity [15:16] without having to rely on a human being to be  there. And hopefully this is going to work. So [15:23] think about putting in a place where  there's not a business model for a human. [15:26] Where's your high intent low conversion and  where's the disproportionate weight of people [15:31] there that need more help? And then again where  buyers need more help than they would have before. [15:36] This is HubSpot which I just talked about but  here's some things they had. Their average [15:40] talk time was 7 minutes from the user, 25%  increase in revenue, 70% free trials. Again, [15:46] the the the AI here took them through from  free trial to buying to onboarding in one flow, [15:51] which was again kind of crazy. Uh, Zoom Info is  a recent customer. They've joined us as well. [15:57] They had a 16x ROI in 5 months. And what they  did is they didn't change anything except for [16:03] put the superhuman behind the form. So  when someone has uh the conversation [16:08] then right behind the form and they and they  fill it out for a sales call the question is [16:12] do you want to have a demo now? They say yes and  it goes into this really deep conversation and [16:16] the results from this are crazy to where they're  getting massive ROI very quickly by just giving [16:20] the option of saying do you want to do more now  versus having to wait for someone or a meeting. [16:27] Um this is Altrix. I just want to show you this  real fast because I I love showing this example. [16:30] So Altrix who came in um they used three main  superhumans. I think it might be four now, [16:34] but May three. There's Annie, who's on their  website. There's Adam, who is their um uh um [16:42] CSM onboarding, and there's Atlas, which is  the ride along that joins calls live and helps [16:47] support the sales rep on the call with technical  questions, which is kind of fun. And here's some [16:51] other like I I can't always share people's uh  companies numbers, so I use them anonymously, [16:56] but this gives you some ideas of other places we  work with based on the size and some other things. [17:02] So, if the question is like who in the heck would  ever want to do this, th this list gives you an [17:06] idea of who's doing this now. Um, this isn't  some theory, this isn't something we're thinking [17:10] it could work. This is working and it's working  really well. And the goal again is to think about [17:15] how can we optimize the buyer's experience so you  get the best of both worlds of when you need the [17:20] best of humanity to build the relationship and  to have that EQ and how can we get the answers [17:25] they need to beforehand so they can move forward  through the process faster. We would love to show [17:30] you more if anyone's interested. Um, please  go to booth 853. We also have tickets for the [17:34] thing tonight. So, if you want some, please come  get me. I'm happy to give you some. Otherwise, [17:38] just want to say thank you for being here and  happy to take any questions if anyone has any.