Advisor Labs' Weidemann Sees AI Spend Wasted Without Executive Push
Ai4
Enterprise AI investments fail to generate clear financial returns because most organizations drop tools into legacy processes without executive leadership or operational redesign.
## Background
This talk was delivered by Chris Weidemann, CAIO and Co-Founder of Advisor Labs, published on YouTube by Ai4 on August 19, 2026. The session explores why widespread enterprise software purchases often fail to deliver real AI value, outlining the gap between individual adoption and true organizational capability.
## TL;DR
Despite high enterprise spend on AI tools, most organizations fail to achieve meaningful business results due to outdated processes, talent gaps, and unmanaged costs. High-performing companies differentiate themselves through executive ownership, AI-first process redesign, and deliberate capacity reallocation. Chris Weidemann outlines five key drivers for success, including establishing AI Enabler roles, protecting proprietary data, targeting high-ROI repetitive workflows, and tracking maturity scores as a proxy for financial return.
## Atmosphere & style
The presentation carries a grounded, practical, and candid tone, opening with lighthearted banter about conference growth and business partners before diving into stark industry metrics. Weidemann communicates with authoritative expertise born from over a decade in machine learning, blending clear macro data with memorable analogies like modern airport design. The talk feels highly actionable and realistic, guiding enterprise leaders away from superficial software hype toward strategic operational discipline.
## Key takeaways
- [00:04](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=4s) Enterprises are heavily investing in AI tools, but very few are achieving true enterprise-level results due to a lack of overall AI capability and strategic integration.
- [02:28](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=148s) A major root cause of poor performance is structural stagnation: 84% of organizations drop AI tools into existing workflows without redesigning roles or business operations.
- [02:28](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=148s) Talent remains a critical issue; even among companies that feel they have secured sufficient AI talent, only about 20% express genuine confidence in their technical capabilities.
- [02:28](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=148s) 80% of enterprise AI budgets are overspent, driven by unpredictable pricing models like API usage fees, fine-tuning costs, and unmanaged shadow AI.
- [06:37](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=397s) Value distribution is heavily skewed toward a small minority of top performers, with the top 20% of enterprises making AI investments capturing roughly 75% of the total market value.
- [06:37](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=397s) Active CEO involvement is a major differentiator: companies where the CEO spends 8+ hours per week on AI are 3 times more likely to be top performers and see higher strategy confidence and realized business value.
- [09:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=567s) Although AI deployment can free up to 8 hours per week for frontline workers, 40% of organizations fail to reallocate this reclaimed capacity toward strategic initiatives, turning efficiency gains into idle time.
- [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) To capture real value, enterprises should designate 'AI Enablers' who combine business domain expertise with AI knowledge to upskill subject matter experts across departments.
- [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) Proprietary business data should be protected as a critical strategic asset rather than casually handed over to third-party models.
- [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) Injecting AI into isolated steps fails; organizations must perform end-to-end process redesigns to prevent simply moving operational bottlenecks downstream.
- [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) The fastest, highest-ROI returns come from targeting repetitive back-office tasks that follow fixed schedules, have standardized methods, and take under five minutes to execute.
- [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) Because direct financial ROI can be hard to measure initially, tracking progress through a numerical AI maturity framework serves as a reliable proxy for ROI to present to executive boards.
## Detailed walkthrough
### [00:04](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=4s) Opening Remarks and the Reality of AI Spending
Chris Weidemann, Chief AI Officer and Co-Founder of Advisor Labs, opens his talk at Ai4 with lighthearted banter, looking back at his business partners in the room and joking about handing them the projector to see if they would step up and deliver the presentation. Observing how dramatically the conference has expanded since previous years, Weidemann highlights his own background—having spent over two decades in progressive technologies and working directly in AI and machine learning since 2012. By 2022, his team made a deliberate pivot to focus exclusively on helping enterprises build, deploy, and capture value from AI.
Addressing the core challenge facing the audience, Weidemann acknowledges that while virtually every enterprise recognizes the necessity of investing in AI, most leadership teams remain deeply uncertain about what specific steps to take, which initiatives will succeed, and what will ultimately fail. The overarching objective of Advisor Labs is to help clients establish a sound AI strategy and navigate critical operational decisions until the client organization can independently sustain momentum.
- Advisor Labs focused 100% of its business on enterprise AI and machine learning deployment in 2022.
- "Everyone knows you need to make an investment in AI, but you're not really sure what you should be doing, how to do it, what's going to produce good results, what's going to fail."
---
[02:28](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=148s) Industry Disconnects: Phantom Value, Talent Gaps, and Overspending
Setting the stage for practical recommendations, Weidemann reviews recent macro data reflecting the state of enterprise AI implementation. A major root cause of poor performance is structural stagnation: 84% of organizations have failed to redesign their roles, business models, or processes around AI, opting instead to simply drop AI tools into existing workflows. Talent acquisition remains a paramount obstacle, and even among companies that feel they have secured sufficient talent, only about 20% express genuine confidence in their technical capabilities.
Furthermore, early optimism regarding ROI is beginning to fade into what Weidemann terms "phantom value." While 42% of enterprises reported at the start of the year that their AI investments were yielding tangible value, mid-year Q2 data indicates that enthusiasm is waning as companies struggle to trace clear financial returns. On the fiscal side, 80% of organizations have exceeded their AI budgets. Part of this stems from unfamiliar cost structures—whereas traditional software spending relies on predictable per-seat SaaS licensing (such as GitHub Copilot or Microsoft 365 Copilot at $30 per month), AI introduces volatile variables like API usage fees, model fine-tuning, and unmanaged "shadow AI."
- 84% of enterprises have dropped AI into old processes without redesigning roles or business operations.
- Only 20% of talent within enterprises is perceived by leadership as truly knowing what they are doing with AI.
- 80% of organizations have overspent their AI budgets, driven by unexpected API usage fees, fine-tuning costs, and shadow AI.
- Finance and finance operations departments are managing budgets slightly better than other business units, but remain roughly 30% over budget overall.
---
[06:37](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=397s) What AI Winners Do Differently: Leadership and Capital Distribution
The distribution of value in enterprise AI is heavily skewed toward a small minority of top performers. Among companies achieving legitimate breakthrough results, 20% of enterprises are capturing roughly 75% of the total value generated across the market. These winning organizations share distinct operational traits: they are three times more likely to approach business problems from an "AI-first" design philosophy rather than attempting retrofitted integrations.
Executive ownership serves as a decisive differentiator in separating successful AI implementations from stalled initiatives. When a Chief Executive Officer actively devotes eight or more hours per week to AI initiatives, the organization is three times more likely to belong to the top-tier "winners group." CEO engagement directly correlates with a 38% increase in strategy confidence across the team, a 36% increase in realized business value, and higher overall financial return. Despite these overwhelming metrics, only one in four enterprises currently has an executive team member actively participating in AI execution on a weekly basis.
- The top 20% of enterprises making significant AI investments generate approximately 75% of the total market value.
- High-performing companies are 3 times more likely to apply an AI-first perspective when evaluating operational problems.
- CEOs spending 8+ hours per week on AI boost strategy confidence by 38% and overall business value by 36%.
- Only 25% (one out of four) of organizations have an executive engaged in AI processes beyond minimal weekly check-ins.
---
[09:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=567s) Reclaiming Capacity vs. Creating Genuine Value
For organizations executing AI strategies correctly, productivity gains are undeniable—data confirms that frontline workers can regain up to a full working day (8 hours) per week. However, Weidemann cautions that reclaimed time does not automatically translate into business value unless leadership actively directs how that capacity is deployed.
Current metrics show that only about 60% of organizations proactively reallocate this newly created capacity toward strategic initiatives. Among individual workers who gain back time without explicit instructions from management, only half take the initiative to find productive tasks on their own. Without deliberate organizational planning, operational efficiency gains degenerate into idle time rather than enterprise growth.
- Successful AI deployment yields up to 1 full day per week returned from frontline workers.
- Only 60% of organizations actively reallocate freed-up employee hours to strategic tasks.
- Only 50% of self-directed employees proactively find productive work when given back unassigned time.
---
[10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) Five Key Drivers for AI Capability and ROI
To help organizations join the high-performing "winners group," Weidemann outlines five concrete drivers that unlock sustainable enterprise AI capability:
#### 1. The "AI Enabler" Role
Enterprises should designate dedicated "AI Enablers"—technically savvy individuals who understand core business operations and are passionate about AI. Their job is to upskill and empower the vital subject matter experts (SMEs) and operational "lynchpins" across departments.
#### 2. Protecting Data as an Enterprise Asset
Organizations must evaluate whether putting proprietary historical data into third-party models dilutes their competitive advantage. As private equity and board-level discussions increasingly reflect, proprietary business decision data carries long-term balance sheet value that should not be given away casually.
#### 3. End-to-End Process Redesign (The Airport Analogy)
Injecting AI into isolated steps of a workflow fails because it merely shifts operational bottlenecks downstream. Illustrating this point, Weidemann compares naive AI integration to software development (where 10x code generation by engineers simply overwhelms review teams) and modern airport security:
> "Airports are designed for us to show up two hours in advance... even though we've added technology, the airport itself is built around the idea of a 2-hour time span. So, if we were to think, how do we rebuild the airport process to be AI first?"
Real enterprise value requires re-engineering entire end-to-end processes around intelligence from the ground up, questioning whether human intervention is even required at specific decision points.
#### 4. Targeting High-ROI Low-Hanging Fruit
Organizations seeking immediate, practical wins should evaluate back-office tasks using three qualifying criteria:
1. Does the task occur on a fixed, predictable schedule (daily, weekly, monthly)?
2. Would two different employees perform the task in roughly the same way?
3. Does the task take less than five minutes to complete individually?
While repetitive administrative tasks lack flashiness, automating these low-hanging workflows yields the fastest, highest-ROI returns.
#### 5. Measuring Maturity as a Proxy for ROI
Calculating precise financial ROI for AI can be elusive when outputs do not map directly to revenue dollars or clear FTE reductions. Weidemann advises leadership teams to establish a quantitative baseline using an AI maturity framework. By scoring organizational maturity periodically (e.g., tracking a movement from a 2.67 baseline in August 2026 to a 3.1 in February 2027), leaders can present executive boards with a concrete, numerical measure of progress and capability growth.
- "If you take this one thing back and you have this one thing for your executive team, I think this is value enough for the week."
- Advisor Labs provides an AI maturity assessment via QR code and hosts discussions at Booth 1125 in the Ai4 exhibit hall.
## Notable quotes
> [00:04](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=4s) Everyone knows you need to make an investment in AI, but you're not really sure what you should be doing, how to do it, what's going to produce good results, what's going to fail. — Chris Weidemann
> [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) Airports are designed for us to show up two hours in advance... even though we've added technology, the airport itself is built around the idea of a 2-hour time span. So, if we were to think, how do we rebuild the airport process to be AI first? — Chris Weidemann
> [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) If you take this one thing back and you have this one thing for your executive team, I think this is value enough for the week. — Chris Weidemann
## Mentioned
Chris Weidemann, Advisor Labs, Ai4, GitHub Copilot, Microsoft 365 Copilot
## Video description
Chris Weidemann - CAIO & Co-Founder, Advisor Labs
Every enterprise bought AI tools this year. Almost none got enterprise results. The pressure to show real value from AI has never been higher, and buying more software is not the answer. This talk will explore why AI wins for individuals rarely become wins for the enterprise, and what the organizations breaking through do differently. You will leave with a clear picture of the gap between AI adoption and AI capability, and practical steps to start closing it. If you are responsible for making AI deliver, start here.
Chris Weidemann of Advisor Labs argues that enterprise AI initiatives frequently stall into budget overruns and 'phantom value' because organizations drop tools into legacy workflows without restructuring processes. Capturing durable returns requires active executive leadership, end-to-end workflow re-engineering, and deliberate management of reclaimed labor capacity.
The Drivers of AI Overspending and Disillusionment (2:28)
Weidemann observes a widening disconnect between enterprise AI investment and actual business return. According to industry data cited in the presentation, 84% of organizations drop AI tools directly into existing processes without redesigning roles or business models, while only 20% of internal technical talent is viewed by leadership as genuinely capable. Consequently, initial enthusiasm is waning into what Weidemann terms 'phantom value': while 42% of enterprises reported tangible returns at the start of the year, mid-year Q2 figures show fading sentiment as organizations struggle to trace clear financial impacts.
Compounding this stagnation, 80% of organizations have exceeded their AI budgets. Traditional software reliance on predictable per-seat SaaS licensing—such as $30 monthly seats for tools like GitHub Copilot or Microsoft 365 Copilot—contrasts sharply with the variable economics of AI. Budget overages are largely driven by unexpected API usage fees, model fine-tuning costs, and unmanaged 'shadow AI,' leaving even finance departments roughly 30% over budget on average.
The Operational Divide Separating AI Winners (6:37)
The economic value generated by enterprise AI is heavily skewed toward a small minority: the top 20% of investing organizations capture approximately 75% of the total value generated across the market. High-performing organizations distinguish themselves through methodology, being three times more likely to apply an 'AI-first' design perspective to operational problems rather than retrofitting tools into legacy systems.
Executive ownership serves as a decisive differentiator. When a Chief Executive Officer actively devotes eight or more hours per week to AI initiatives, the organization is three times more likely to belong to the top-tier winners group. This executive commitment correlates with a 38% increase in strategic confidence and a 36% rise in realized business value. Despite these outcomes, executive involvement remains an industry exception, with only 25% of enterprises reporting weekly executive participation in AI execution.
Reclaiming Capacity Demands Deliberate Process Redesign (9:27)
When deployed effectively, AI can return up to one full working day (eight hours) per week to frontline staff. However, Weidemann emphasizes that reclaimed time does not automatically translate into enterprise value unless management actively directs where that labor is applied. Current data indicates that only 60% of organizations systematically reallocate freed-up employee hours to strategic initiatives. Left without explicit management guidance, only 50% of self-directed workers independently locate productive tasks, risking the loss of efficiency gains to idle time.
Moreover, inserting AI into isolated tasks fails because it merely shifts operational bottlenecks downstream. Comparing naive integration to software development—where generating ten times more code simply overwhelms human reviewers—and airport security built around a fixed two-hour arrival buffer, Weidemann argues that real transformation requires questioning whether human checkpoints are necessary and redesigning entire end-to-end workflows around intelligence.
Concrete Frameworks for Building Long-Term Capability (10:27)
To build durable capability, Weidemann highlights key operational levers. Organizations should designate internal 'AI Enablers'—technically competent operators who upskill vital subject matter experts and departmental lynchpins. Enterprises must also safeguard proprietary historical decision data as a balance-sheet asset, avoiding the casual transfer of competitive advantages to third-party model providers. For immediate wins, organizations should target high-ROI, low-hanging administrative workflows that occur on a predictable schedule, follow consistent steps across different workers, and take less than five minutes individually.
Because financial ROI and head-count reductions can be difficult to quantify directly, Weidemann recommends tracking progress through a numerical AI maturity framework. Documenting periodic maturity scores—such as moving from a 2.67 baseline in August 2026 to 3.1 in February 2027—provides executive boards with an objective, demonstrable proxy for capability growth.
Weidemann concludes that enterprise transformation depends on transitioning from fragmented tool adoption to systematic capability building. However, the framework leaves open how leadership should address workforce impacts if operational time savings cannot be strategically redeployed, relying instead on internal upskilling and maturity scoring to justify sustained investment.
## Background
本内容由 Ai4 于 2026 年 8 月 19 日在 YouTube 平台发布，主讲人为 Advisor Labs 首席 AI 官兼联合创始人 Chris Weidemann。视频深入探讨了为什么企业仅仅采购 AI 软件往往难以取得实质性商业成果，并剖析了从个体效率提升跨越到构建企业级 AI 核心能力的切实路径与策略。
## TL;DR
演讲指出，绝大多数企业陷入了盲目采购工具、遭遇预算超支以及面临人才与价值幻觉的困境，全行业仅有 20% 的赢家捕获了近四分之三的实际商业价值。要实现真正的价值突破，企业需要 CEO 深度参与战略决策，并设立兼具业务与技术敏锐度的“AI 赋能者”角色来带动核心业务骨干。同时，企业应避免孤立的局部修补，转向以“AI 优先”思维重构端到端业务流程，并主动规划员工释放出的工时。最后，演讲提供了识别后台高频速赢场景的“三问评估法”，以及通过 AI 成熟度模型量化长期投资回报率的务实解决方案。
## Atmosphere & style
整场演讲务实、干练且直击痛点，开场以轻松自嘲打破坚冰，随后迅速切入扎实的数据与行业现实分析。主讲人风格沉稳专业，摒弃了空洞的行业炒作，以顾问和实践者的视角层层剖析企业落地难点，展现出一种冷静反思与建设性指导并存的专业氛围。听众能清晰感受到紧贴企业管理一线的紧迫感与落地方法论的启发感。
## Key takeaways
- [00:04](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=4s) 企业 AI 转型的目标应是从购买工具转向构建自主可控的 AI 体系能力（AI Capability），而非持续依赖外部顾问或盲目采购软件。
- [02:34](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=154s) 84% 的企业尚未围绕 AI 重新设计业务流程与岗位职责；在自认拥有 AI 人才的企业中，仅有 20% 对自身团队的落地胜任力抱有信心。
- [02:34](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=154s) 许多企业在初期感知到的投资回报只是一种“幻觉价值（Phantom Value）”，随着实际业务数据的出炉，最初的乐观情绪往往迅速消退。
- [05:05](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=305s) 高达 80% 的企业 AI 预算超支，根源在于动态 API 调用、模型微调算力以及影子 AI 等隐蔽且波动的成本打破了传统固定席位费的监控体系。
- [05:05](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=305s) AI 投资收益呈现高度集中的二八定律，仅约 20% 具备“AI 原生（AI-First）”重构思维的领跑者捕获了全行业约 74% 至 75% 的实际商业价值。
- [08:01](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=481s) 高管深度参与是成功的关键指标：CEO 每周投入 8 小时以上在 AI 战略上的企业，成为高回报赢家的概率提升 3 倍，但目前仅 25% 的企业做到了这一点。
- [08:01](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=481s) “工时释放”并不等同于“价值创造”，约 40% 的企业在 AI 节省时间后未做任务再分配规划，导致释放出的工时有半数被闲置或浪费。
- [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) 打破落地隔阂的有效机制是设立“AI 赋能者（AI Enabler）”角色，通过兼具业务理解与技术敏锐度的人才结对赋能业务核心骨干（SMEs）。
- [12:06](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=726s) 企业必须重视自身业务决策数据资产的流失风险，避免在未经保护的情况下让第三方模型供应商无偿获取并提炼商业价值。
- [12:06](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=726s) 局部孤立的工具提速只会将瓶颈推移至下游（如研发提速导致代码审查严重积压），必须以 AI 优先思维对端到端业务流程进行彻底重构。
- [15:39](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=939s) 识别后台运营中高 ROI 速赢场景的“三问评估法”：任务是否按固定周期发生、多人操作是否高度一致、单次执行是否少于 5 分钟。
- [16:53](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=1013s) 在复杂场景难以直接测算财务收益时，引入“AI 成熟度评估模型”进行阶段性基准与复测打分，是向董事会量化汇报 AI 进展的有效路径。
## Detailed walkthrough
### [00:04](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=4s) 开场与背景：从 AI 浪潮到企业的现实困境
演讲者 Chris Weidemann 以一段轻松的玩笑开场，调侃自己站在台上时总有种想把放映机丢给台下的合伙人、自己偷偷溜走的冲动。他回顾了自身参与 Ai4 大会的历程，感叹整个 AI 生态和大会规模在过去几年间呈现出爆发式增长。
在个人背景方面，Chris 拥有超过二十年的前沿科技从业经历，自 2012 年起便深耕 AI 和机器学习领域。随着行业关注度的提升，团队在 2022 年决定破釜沉舟，将全部业务聚焦于 AI 领域，致力于帮助企业构建、部署并从 AI 中捕获真实价值。
Chris 指出，当前大多数企业面临的普遍困局在于：几乎所有管理者都清楚必须对 AI 进行战略投资，却极少有人明确知晓具体该做什么、如何落地、哪些项目能产生效益而哪些注定会失败。Advisor Labs 衡量自身成功的标准非常务实——在介入企业、理顺战略与决策框架后，能够让企业团队自主掌控并运转 AI 体系，顾问团队即可平稳抽身离开。
```mermaid
flowchart LR
accTitle: 企业 AI 落地面临的困境与 Advisor Labs 的定位
A["传统业务困境<br/>(盲目买工具/不知如何落地)"] --> B["Advisor Labs 介入<br/>(梳理战略/决策/落地路径)"]
B --> C["企业自主掌控能力<br/>(团队具备 AI Capability)"]
C --> D["顾问顺利抽身<br/>(产生可持续企业价值)"]
```
---
### [02:34](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=154s) 行业现状：AI 投资中的人才短缺与“幻觉价值”
在有限的演讲时间内，Chris 梳理了当前行业最新、最切实的几组关键数据，以此揭示 AI 落地过程中的结构性矛盾：
84% 的企业尚未围绕 AI 重新设计业务流程、岗位职责与组织架构，绝大多数只是简单地将 AI 工具生硬地塞进现有的既有流程中。当被问及实施 AI 的最大障碍时，企业几乎一致认为是“难以寻觅合适的人才”；而在自认为拥有相关人才的企业中，仅有约 20% 对自身团队的技术与落地能力抱有充分信心。
```mermaid
flowchart TD
accTitle: 企业 AI 落地中的痛点与幻觉价值
subgraph 组织与人才痛点
T1["84% 企业未围绕 AI 重构业务与岗位"]
T2["最大阻碍：顶尖 AI 人才极度匮乏"]
T3["仅 20% 人才被认为具备真正胜任力"]
end
subgraph 价值感知的落差
V1["Q1：42% 企业自认获得投资回报"] --> V2["Q2-Q3：实际数据出炉，信心动摇"]
V2 --> V3["普遍陷入'虚假/幻觉价值 (Phantom Value)'"]
end
```
这种能力断层直接导致了企业价值感知的剧烈落差：在年初，约 42% 的企业管理层信心满满，认为自身的 AI 投资正在带来实质性的企业价值；然而进入第三季度、面对第二季度的实际业务数据后，这种乐观情绪迅速消退，企业开始怀疑这些投资是否真的见效。年初所谓的繁荣在很大程度上只是一种“幻觉价值”（phantom value），业内绝大多数组织至今仍未建立起稳固的价值回报信心。
---
### [05:05](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=305s) 预算失控与赢家法则：80/20 法则与 AI 原生思维
在财务与成本控制维度，AI 项目的预算超支已成为行业常态。数据显示，高达 80%（即五分之四）的企业超出了原定的 AI 预算。即便是表现相对最好的财务与运营部门，实际支出也普遍超支约 30%。
预算失控的根源在于成本构成的隐蔽性与不可预测性。传统 IT 采购习惯于清晰的软件许可模式（例如每人每月 30 美元的 Copilot 席位费，在账单上一目了然）；但在 AI 时代，企业必须面对 API 调用账单、模型微调（fine-tuning）费用以及大量未受监管的“影子 AI”（Shadow AI）支出。这些零散且动态波动的成本结构，是传统财务监控体系极难捕捉的。
```mermaid
flowchart LR
accTitle: 传统 IT 成本与现代 AI 成本结构的对比
subgraph 传统软件支出
S1["按席位订阅 (如 $30/月)"]
S2["固定 IT 账单与清晰科目"]
end
subgraph 现代 AI 动态支出
A1["API 动态调用计费"]
A2["模型微调与训练算力"]
A3["影子 AI 零散支出"]
end
```
在严峻的现实中，少数“AI 赢家”展现出了截然不同的运作模式：
- **AI 原生思维（AI-First）**：成功企业在审视业务痛点时，采用“AI 优先”视角的概率是普通企业的 3 倍。他们不再思考“如何把 AI 拼接到旧业务上”，而是思考“如何以 AI 为核心重构整个业务流程”。
- **价值的高度集中（二八定律）**：在所有进行大规模 AI 投资的企业中，仅有约 20% 的领跑者攫取了全行业约 74% 至 75% 的实际商业价值，AI 投资绝非“只要投钱就必有回报”的确定性游戏。
---
### [08:01](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=481s) 关键胜负手：CEO 的深度参与与工时释放陷阱
数据表明，高管层的深度参与是企业能否跻身 AI 赢家行列的最强指标。当一家企业的 CEO 每周在 AI 战略和举措上投入 8 小时以上时，该企业成为高回报赢家（捕获 74% 行业价值）的概率将提升 3 倍。
拥有 CEO 作为直接负责人的企业展现出了全方位的优势：全员对战略方向的信心提升 38%，业务价值产出提升 36%，投资回报率（ROI）也显著增高。然而，现实情况十分严峻：目前全行业仅有四分之一（25%）的企业真正做到了让核心高管每周深度参与 AI 决策。
```mermaid
flowchart TD
accTitle: 释放工时向业务价值转化的漏斗
A["实施优秀 AI 举措：一线员工每周节省 1 整天 (8小时)"] --> B{"时间再分配机制"}
B -- "约 60%" --> C["企业主动重新分配该工时"]
B -- "约 40%" --> D["企业未作明确安排"]
D --> E{"员工自主性"}
E -- "仅 50%" --> F["员工主动寻找高价值任务"]
E -- "另 50%" --> G["闲置/无效消耗 (未转化为商业价值)"]
```
在效益转化层面，走在正确路径上的企业能够为一线员工每周节省整整一天（8 小时）的工作时间。但“节省时间”并不等同于“产生价值”：
- 仅有约 60% 的企业会主动规划并将释放出的时间重新分配到高价值任务中。
- 在剩下未获得明确指令的员工中，仅有一半人会主动寻找有意义的新工作，其余人员释放出的时间则被白白闲置或浪费。
---
### [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) 核心驱动力之一：设立“AI 赋能者 (AI Enabler)”角色
为了打破技术与业务之间的隔阂，Chris 提出了企业跻身赢家行列的关键抓手——在组织内部设立“AI 赋能者”（AI Enabler）角色。
AI 赋能者需要兼具三大特质：深刻理解业务运作逻辑、具备扎实的技术敏锐度、对 AI 充满极高热情。该角色的核心使命并非单打独斗，而是去赋能组织内部那些至关重要的业务骨干与领域专家（SMEs / Subject Matter Experts）——即各部门里人人都去请教、维系业务运转的核心“中流砥柱”（lynchpins）。
```mermaid
flowchart LR
accTitle: AI 赋能者的核心联动机制
AE["AI 赋能者 (AI Enabler)<br/>• 懂业务<br/>• 懂技术<br/>• 懂 AI 落地"] -->|赋能与加速培训| SME["业务核心骨干 / 领域专家 (SMEs)<br/>• 业务中流砥柱 (Lynchpins)<br/>• 掌握实际痛点"]
SME -->|结合 AI 工具深度重构| V["爆发式产生业务 ROI 与企业能力"]
```
通过让 AI 赋能者与业务核心骨干结对，企业能以极快的速度让真正懂业务的人掌握 AI 核心能力，从而将个体的工具效率迅速放大为组织级的核心能力与显著的 ROI 回报。紧接着，Chris 进一步引出下一个关键驱动要素：企业的数据资产价值。
### [12:06](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=726s) 数据资产与端到端的业务流程重构
企业在将数据输入大模型之前，必须先进行深度自省：这些数据究竟是对当前或历史业务决策产生核心指导价值的资产，还是正在被无偿拱手让给第三方供应商？如果第三方能够借由企业的历史决策数据提炼出属于他们自己的商业价值，企业就面临着巨大的价值流失。随着私募股权投资机构（Private Equity）开始关注资产负债表中的数据价值，企业数据的保护与变现已成为不可忽视的高管层对话焦点。
```mermaid
flowchart TD
accTitle: 从局部工具注入到端到端重构
subgraph Local["局部修补：转移瓶颈"]
direction LR
A["研发编码"] -->|"AI 提速 10x"| B["代码审查 Backlog 积压"]
B --> C["产品交付整体停滞"]
end
subgraph EndToEnd["端到端重构：以 AI 为先"]
direction LR
D["起点：业务输入"] --> E["重构：判断节点由 AI 还是人类负责？"]
E --> F["终点：产品/商业价值输出"]
end
```
在将 AI 融入业务时，最大的误区在于仅把 AI 机械地“注入”到某一个孤立环节，而忽视了整体流程。企业应当审视从起点到终点的整个链条，提出根本性的重构问题：“如果从第一天起就拥有 AI，我们会如何重新设计这个流程？”例如在软件工程领域，给工程师配备 AI 工具后，代码产出速度提升了 10 倍，但由于下游的代码审查（Code Review）流程并未重新设计，庞大的待审积压让整体产品交付速度丝毫没有提升，仅仅是将瓶颈向下游推移。这与现代机场的设计逻辑如出一辙：尽管几十年来引入了大量安检与值机技术，但因为整套系统依然围绕“提前 2 小时到机场”的旧流程运作，旅客依然无法在 30 分钟内完成登机。真正的价值突破，唯有彻底以“AI 优先”的理念重构端到端流程才能实现。
- **核心反思**：不要问“哪里可以塞入 AI”，而要问“在 AI 赋能的前提下，这个业务决策还需要人类参与还是交给算法”。
- **局部优化的陷阱**：单纯提升上游 10 倍产出，若不重构下游协同，只会将瓶颈转移至审核环节，无法转化为企业级成果。
---
### [15:39](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=939s) 识别速赢场景：三问评估法
对于寻求快速见效的企业，寻找“低垂的果实”（Low-hanging Fruit）有着非常明确的筛选标准。后台运营中那些虽然看起来“不性感”但高度重复的日常事务，往往具备极高的投入产出比（ROI）。企业员工只需回答三个简单的问题，就能迅速在自身工作中定位出 5 到 10 个绝佳的自动化落地点：
```mermaid
flowchart LR
accTitle: 识别 AI 低垂果实的三个核心问题
Q1["1. 是否按固定周期发生？<br/>(每天/每周/每月)"] --> Q2["2. 多人操作方式是否高度一致？<br/>(流程标准化、少偏差)"]
Q2 --> Q3["3. 单次执行是否少于 5 分钟？<br/>(高频碎片化任务)"]
Q3 --> Target["极高优先级 AI 自动化场景<br/>(后台运营高 ROI)"]
```
- **发生周期明确**：任务是否按日程表固定发生（例如每天、每周一上午或每月初）？
- **流程高度一致**：如果安排不同办公室的两位员工来做这件事，两人的操作方式是否基本一致？
- **耗时极短**：单次任务的耗时是否在 5 分钟以内？
---
### [16:53](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=1013s) 用成熟度模型破解企业 ROI 度量难题
对于“如何衡量 AI 的投资回报率”这一行业普遍痛点，除少数能直接折算为金额或明确节省 20% 全职人力（FTE）的简单任务外，大多数业务场景的量化极其复杂。为了破除无法向董事会或管理层交代的困境，企业应当引入“AI 成熟度评估”（Maturity Assessment）作为基准衡量指标。
```mermaid
flowchart LR
accTitle: 基于成熟度评估的 ROI 衡量路径
A["基准测试<br/>(如 2026 年 8 月: 2.67 分)"] --> B["推行 AI 战略落地与流程重构"]
B --> C["复测评估<br/>(如 2027 年 2 月: 3.10 分)"]
C --> D["向董事会/高管汇报：<br/>成熟度量化提升 +0.43"]
```
企业可以在落地初期通过成熟度评测获取一个初始基准分数（例如打出 2.67 分），在经过 3 到 6 个月的落地推进后再次复测。当评估分数上升至 3.1 时，团队便能向高管层提供一个扎实、具象的成果指标——“过去一年的投入使企业的 AI 成熟度提升了 0.43”。在绝大多数企业依然拿不出任何 ROI 数据的现状下，用成熟度成长来量化回报是最行之有效的方法。演讲最后，Chris Weidemann 邀请现场观众在会后及 1125 号展位继续交流探讨。
## Notable quotes
> [00:04](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=4s) 几乎所有管理者都清楚必须对 AI 进行战略投资，却极少有人明确知晓具体该做什么、如何落地、哪些项目能产生效益而哪些注定会失败。 — Chris Weidemann
> [02:34](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=154s) 84% 的企业尚未围绕 AI 重新设计业务流程、岗位职责与组织架构，绝大多数只是简单地将 AI 工具生硬地塞进现有的既有流程中。 — Chris Weidemann
> [05:05](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=305s) 高达 80% 的企业超出了原定的 AI 预算。即便是表现相对最好的财务与运营部门，实际支出也普遍超支约 30%。 — Chris Weidemann
> [05:05](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=305s) 在所有进行大规模 AI 投资的企业中，仅有约 20% 的领跑者攫取了全行业约 74% 至 75% 的实际商业价值。 — Chris Weidemann
> [08:01](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=481s) 当一家企业的 CEO 每周在 AI 战略和举措上投入 8 小时以上时，该企业成为高回报赢家的概率将提升 3 倍。 — Chris Weidemann
> [08:01](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=481s) 节省时间并不等同于产生价值。 — Chris Weidemann
> [10:27](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=627s) AI 赋能者的核心使命并非单打独斗，而是去赋能组织内部那些至关重要的业务骨干与领域专家。 — Chris Weidemann
> [12:06](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=726s) 如果从第一天起就拥有 AI，我们会如何重新设计这个流程？ — Chris Weidemann
> [12:06](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=726s) 不要问‘哪里可以塞入 AI’，而要问‘在 AI 赋能的前提下，这个业务决策还需要人类参与还是交给算法’。 — Chris Weidemann
> [12:06](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=726s) 单纯提升上游 10 倍产出，若不重构下游协同，只会将瓶颈转移至审核环节，无法转化为企业级成果。 — Chris Weidemann
> [15:39](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=939s) 后台运营中那些虽然看起来‘不性感’但高度重复的日常事务，往往具备极高的投入产出比。 — Chris Weidemann
> [16:53](https://www.youtube.com/watch?v=XAMSbS6CBsY&t=1013s) 在绝大多数企业依然拿不出任何 ROI 数据的现状下，用成熟度成长来量化回报是最行之有效的方法。 — Chris Weidemann
## Mentioned
Chris Weidemann, Advisor Labs, Ai4, Copilot
## Video description
Chris Weidemann - CAIO & Co-Founder, Advisor Labs
Every enterprise bought AI tools this year. Almost none got enterprise results. The pressure to show real value from AI has never been higher, and buying more software is not the answer. This talk will explore why AI wins for individuals rarely become wins for the enterprise, and what the organizations breaking through do differently. You will leave with a clear picture of the gap between AI adoption and AI capability, and practical steps to start closing it. If you are responsible for making AI deliver, start here.
[Source](https://www.youtube.com/watch?v=XAMSbS6CBsY)
Source quotes
[0:04] Chris Weidemann: 几乎所有管理者都清楚必须对 AI 进行战略投资，却极少有人明确知晓具体该做什么、如何落地、哪些项目能产生效益而哪些注定会失败。
[2:34] Chris Weidemann: 84% 的企业尚未围绕 AI 重新设计业务流程、岗位职责与组织架构，绝大多数只是简单地将 AI 工具生硬地塞进现有的既有流程中。
[5:05] Chris Weidemann: 高达 80% 的企业超出了原定的 AI 预算。即便是表现相对最好的财务与运营部门，实际支出也普遍超支约 30%。
[5:05] Chris Weidemann: 在所有进行大规模 AI 投资的企业中，仅有约 20% 的领跑者攫取了全行业约 74% 至 75% 的实际商业价值。
[8:01] Chris Weidemann: 当一家企业的 CEO 每周在 AI 战略和举措上投入 8 小时以上时，该企业成为高回报赢家的概率将提升 3 倍。
[8:01] Chris Weidemann: 节省时间并不等同于产生价值。
[10:27] Chris Weidemann: AI 赋能者的核心使命并非单打独斗，而是去赋能组织内部那些至关重要的业务骨干与领域专家。
[12:06] Chris Weidemann: 如果从第一天起就拥有 AI，我们会如何重新设计这个流程？
[12:06] Chris Weidemann: 不要问‘哪里可以塞入 AI’，而要问‘在 AI 赋能的前提下，这个业务决策还需要人类参与还是交给算法’。
[12:06] Chris Weidemann: 单纯提升上游 10 倍产出，若不重构下游协同，只会将瓶颈转移至审核环节，无法转化为企业级成果。
[15:39] Chris Weidemann: 后台运营中那些虽然看起来‘不性感’但高度重复的日常事务，往往具备极高的投入产出比。
[16:53] Chris Weidemann: 在绝大多数企业依然拿不出任何 ROI 数据的现状下，用成熟度成长来量化回报是最行之有效的方法。
Source quotes
[0:04] Chris Weidemann: Everyone knows you need to make an investment in AI, but you're not really sure what you should be doing, how to do it, what's going to produce good results, what's going to fail.
[10:27] Chris Weidemann: Airports are designed for us to show up two hours in advance... even though we've added technology, the airport itself is built around the idea of a 2-hour time span. So, if we were to think, how do we rebuild the airport process to be AI first?
[10:27] Chris Weidemann: If you take this one thing back and you have this one thing for your executive team, I think this is value enough for the week.
Original transcript
[0:04] appreciate it. You know, all of us kind of have
that devil on our shoulder and I'm standing up
[0:08] here and I see a couple of my business partners
back there and I'm wondering like what happens if
[0:13] I just leave the room and hand them the projector?
Does this does this go okay? Do they get up here
[0:18] and they do a good job? I'm sure they would.
But this is something that we've been planning
[0:21] for since last year. We were here last year,
the year before. This has really grown. Just
[0:26] kind of raise of hands. Who who was here last
year? who participated in AI4 last year. Looks
[0:32] like a couple. Has anyone else noticed just how
much this thing has grown in the last few years?
[0:39] It's much much bigger, which is great. We all love
participating in this conversation. And you know,
[0:46] for me, this is this is fun. I I enjoy this. As
a little bit about me, maybe why, you know, I'm
[0:54] even worth spending some time thinking about. I've
been doing this for a long time. So progressive
[1:00] technologies for more than two decades got into
AI and ML back in 2012 and slowly as the rest of
[1:09] everyone else has found it interesting we've done
more and more of of AI and ML. So much so that in
[1:16] 2022 we decided you know what let's only focus
on this. Let's let's not focus on anything else.
[1:23] let's just focus on helping companies find great
ways to build, deploy, and find value with AI. So,
[1:30] that's what we've been doing. And as an
organization, we really look at the problem that's
[1:36] out there. And what we've heard from many people
like yourself is that everyone knows you need to
[1:42] make an investment in AI, but you're not really
sure what you should be doing, how to do it,
[1:49] what's going to produce good results, what's going
to fail. These are all hard things that we're
[1:55] struggling with. Now, luckily because we've had
an opportunity to work with many organizations,
[2:01] we've kind of gone through those cycles, those
iterations, and so we know what's working. We
[2:07] definitely know what's what's not working. But
we consider ourselves successful when we're
[2:12] able to disengage from a client. We've come
in, we've helped them make those decisions,
[2:17] understand strategy, understand what those good
decisions look like, then they're taking over.
[2:23] they're running with it and we can remove
ourselves from the picture. So that's
[2:28] really what we do. If we talk about what we're
going to talk about here in this this section,
[2:34] we got about 20 minutes, 17 minutes and 1 second
to be precise. Now, uh, and that's not a whole lot
[2:40] of time for us to take a lot of information, the
things that we've learned over the last year and
[2:45] distill it down into one conversation. But that's
what I'm going to try to do. And I'm going to do
[2:50] that. We're going to frame it by talking about
what all of us as an industry know. So I've taken
[2:57] what I think are some of the most relevant bits of
information that's coming out recently about how
[3:03] AI is working or is not working inside of industry
right now. And we're going to use that to set
[3:09] the stage for the five pieces at the end that we
know drive value in all organizations. So what do
[3:18] those data points look like? Great example, 84% of
companies have not redesigned businesses, roles,
[3:28] company processes around AI. Instead, they've just
found opportunities to drop AI into whatever was
[3:36] already there. When you ask companies, what's
the hardest part of AI? Almost exclusively,
[3:43] they will tell you finding good talent. It's
really hard to find people. And then when you
[3:49] talk to the people that feel like, okay, yeah, we
we've we have some of this talent under control,
[3:55] most of them will say only about 20% of those
people, they feel confident that they really know
[4:01] what they're doing. So there's all of these kind
of gaps in the market. Now, there's good things
[4:07] too, right? We know that at the beginning of this
year, enterprises felt a lot of confidence. They
[4:14] knew that the investments that they were making,
42% of them felt like the investments that they
[4:18] were making were producing some type of enterprise
value. So, hey, yeah, we're making good. We're
[4:24] making investments. They're returning something.
But now, as we enter into quarter 3, we've
[4:30] actually seen the numbers from quarter two and
something's happening where that that excitement,
[4:37] all of that interest or, you know, yes, this is
producing value. Now we're hearing back from those
[4:43] same companies and they're saying uh I don't I'm
not sure if this is producing value. So there's a
[4:49] little bit of phantom value there in the beginning
of the year where people were feeling excitement.
[4:53] So there's still most of the people in this room,
most of the organizations that are represented
[4:58] in this room still are yet to feel confident
about the value that they're producing with AI.
[5:05] Now, we know the other hard part with AI as
an industry is the money side. Basically,
[5:13] 80% four out of five organizations have overspent
their AI budgets. Now, we're all really good at
[5:22] overspending budgets, right? It's not just an
AI problem. We we do this with many budgets,
[5:27] but I think 80% is a really big number. And
so, if you look at, okay, who's doing it
[5:32] really well? If you kind of break that down and
look at the departments inside of organizations,
[5:38] you look at finance, finance ops, they are
doing a little bit better than everybody else,
[5:43] but it's still basically 30% over budget in total.
So across the board, we're having a hard time
[5:50] keeping these budgets within the realm of where we
expected them. Now, there's good reason for that,
[5:56] and a lot of it has to do with it's really hard
to see where all that money is getting spent.
[6:02] We're all used to spending money with a license.
Well, yeah, let me turn on co-pilot and I know
[6:06] that's $30 a month. I can track that. Like, I know
where to look for that on a line item. But there
[6:12] are many things that organizations are not used to
tracking. A lot of organizations were not used to
[6:17] paying for an API bill or thinking about tuning
or all of this shadow AI that's out there now.
[6:24] And so we have all these bits and pieces that are
floating out there in the cost structure that we
[6:30] as organizations are not really used to looking
at or looking for. So 80% of us are spending more
[6:37] than we thought we would. Now you know there
there are certainly positive things too that
[6:43] we're getting out of this. We know that of the
winners, the people that are winning with AI,
[6:49] they're doing some things right. If you're winning
with AI, if you consider yourself winning with AI,
[6:55] those people are three times more likely to look
at any problem they're solving from an AI first
[7:02] perspective. So instead of looking at a problem
and saying, "Oh, let me find a way of integrating
[7:06] AI with that problem, let me look at that and how
do I redesign that with an AI first philosophy?
[7:14] How do I build this around AI?" Now that value
again right it's kind of biased and there is a
[7:24] small number of us that are getting the large
share out of the value from these investments.
[7:31] So even out of those winners we know that roughly
20% of the enterprises that are making significant
[7:38] investment in AI those 20% are getting roughly
75% of the value. So, this is not a game where
[7:45] we all play that if we know we put some money
into this, we know we're going to get a return.
[7:51] And that's what makes this hard. That's why all
of us are are one of the reasons why we are here
[7:56] today. We're all trying to figure out how to
do this in a better way, in a new way. One of
[8:01] the indicators that we look for when we go into an
enterprise is how involved is the executive team.
[8:08] We know that if we can look at a CEO's calendar
and he's spending or he or she is spending eight
[8:14] plus hours a week on AI initiatives that they're
part of that winning group. So, we need to get our
[8:21] executives involved. They need to be a part of
this process. We also know that they are three
[8:27] times more likely if they're spending eight
hours a week or more, they're three times more
[8:32] likely to be a part of that AI winners group.
the ones that are excelling, the ones that are
[8:37] getting 74% of the value. And you know, there's
a benefit of having all executives involved,
[8:46] but if you're going to pick an executive, the data
tells us it should be the CEO. If you have the CEO
[8:53] involved as that owner, you're 38% more likely to
have confidence, the entire team having confidence
[8:59] in the strategy that's been selected. You get
more business value out of those investments,
[9:05] 36%, and there's more ROI in return. But here's
the killer. We know that only a fourth of the
[9:13] businesses out there have actually done this.
There only one out of of four actually have an
[9:20] executive member that's a part of this process
in more than just a little bit each week.
[9:27] So if we look at all of these companies that are
making these good decisions, they're doing the
[9:32] right thing, we can expect to get back a full day
a week from frontline workers. The data is now in.
[9:39] We know what the return on these investments are.
And so if you're doing the right things and you're
[9:44] getting those investments, you're part of the
20% that's getting the 74% back. You're getting
[9:50] a day a week back from frontline workers. But it
doesn't end there. right now. What do you do with
[9:56] that time? And we know that only about half,
60% are taking that time and redistributing it
[10:03] somewhere else. And of the people that are getting
that time back and are not being told what to do,
[10:09] only half of those people are proactively finding
something to do on their own. So even though we're
[10:15] getting this capacity back, it doesn't mean that
it's value. Instead, it's just time where people
[10:20] can sit and maybe do something else. We're not
really sure what they're doing. Okay, here's
[10:27] where we're going to talk about the things that
we know are those drivers, those knobs and dials
[10:32] that you need to turn in 27 to be a part of that
AI winners group. So, if I've lost you all, here's
[10:40] where you come back. We know in the same way that
having an executive involved there is this great
[10:48] trick of having a role called an AI enabler. We
find individuals that understand business that are
[10:57] technically savvy are excited about AI and we make
them responsible for accelerating the growth and
[11:05] development of those business thosememes those
subject matter experts the business experts in
[11:11] your organization that truly understand the
business. Now all of you know the people in
[11:17] your organization that are truly vital, the ones
that produce, the the ones that everyone goes
[11:23] to if they have a question, the ones that are,
you know, really the lynch pin in the business.
[11:29] Having an AI enabler as someone that that person
can utilize is a significant significant return
[11:38] on investment. you all of a sudden you give
these vital lynch pins of your business the
[11:44] opportunity to learn and quickly develop all
of the tools that everyone in this room is now
[11:50] understanding the power of. You're giving those
people that power very very quickly. Your data,
[11:58] here's one that many don't think about. Your data
is valuable. As you're making investments in 27,
[12:06] you really should be asking yourself, is this data
that I'm about to put into this model something
[12:13] that helps me as a business or historically helps
me understand something about my business? Am I
[12:20] going to be giving this data away to a third
party that can then take my historical business
[12:26] decisions and turn that into value themselves?
This is the conversation a lot of businesses
[12:32] should be asking themselves right now. and should
I really be giving my business data away to other
[12:36] people? There's a lot of value here that's hard
to unlock today, but will quickly. We're seeing
[12:42] this on on balance sheets as private equity is
getting involved. We're seeing this as as part
[12:46] of many of the conversations. All right, we
I've hinted at this one three times and I'm
[12:52] sure you've heard it other times this week as
well. Maybe one of the biggest dials you can
[12:57] turn is looking at anything you're adding AI to.
Don't just look at the one little piece. Look at
[13:04] it as an entire process. Go to the very beginning
and look at the very end. And don't ask yourself,
[13:10] where can I inject AI? But instead, how would
I do this differently? How would I change this
[13:16] process if I had AI from the very beginning? Would
I have these people involved? Would I have humans
[13:24] involved here or would I have a computer making
a decision here? Those of us that can find those
[13:30] opportunities and the time to re-engineer
processes around intelligence are the ones
[13:36] that are really going to find the value. Now, I I
like sharing the example of software development
[13:42] uh because it's a it's a great one. Software
development, if you're a product company right
[13:47] now, you found that you can get your engineers
to produce 10x more code by giving them tools.
[13:52] And it's happening. Engineers are producing
more code. But across many organizations,
[13:58] many software organizations out there right now,
just because they're producing 10x from a software
[14:04] side doesn't mean that they're producing
10x from a product side. What's happening
[14:08] is they're just moving the bottleneck down the
chain. So now the review team has got a bigger
[14:13] backlog than they ever had before. It's not that
we're producing more as a product. It's just we've
[14:19] moved it downstream. Now we all know this to be
true because we've experienced it in our lives
[14:23] in many different ways. Those of you that have
traveled here this week, many of you probably came
[14:29] through an airport and you've experienced exactly
this. How many of you have heard for more than a
[14:35] decade now that you should arrive 2 hours ahead
of time before your flight, right? Most of us,
[14:41] most hands going up. And that's because that's
what we've been told. Airports are designed for
[14:46] us to show up two hours in advance. Now, there's
a lot of technology that's gone into that process,
[14:52] but is anyone really showing up in in 1 hour
or 30 minutes before the time on on average?
[14:59] Not really. We're all built around this idea of
it taking 2 hours to get through an airport. And
[15:04] that's because even though we've added technology,
the airport itself is built around the idea of a
[15:10] 2-hour time span. So, if we were to think, how
do we rebuild the airport process to be AI first?
[15:17] you know, maybe we could get this down ways. But
this is the philosophy you've got to take back to
[15:22] your businesses and ask yourself, okay, we want
to integrate, we want to add intelligence into
[15:27] this part of our business, but we don't just
want to add it. We want to redesign that part
[15:32] of the business to leverage intelligence.
That's the question you should be asking.
[15:39] Okay, this is probably another one you've heard
this week, but I'm going to I'm going to reinforce
[15:43] it because it is the lowest of hanging fruit
that many companies don't follow. There's a
[15:50] very simple way of figuring out if you have really
great lowhanging fruit opportunities with AI. You
[15:57] just ask yourself these three questions. Does
it happen on a schedule? Am I doing this daily,
[16:04] weekly, every other week, once a month? Can I name
how frequently this thing is happening? If two or
[16:11] more people were to do this same thing, would
they do it in roughly the same way? You know,
[16:16] someone in this office does it this way,
someone in that office does it just about
[16:20] the same way? That's a great indicator. And
then does it take less than five minutes? Now,
[16:26] I'm sure if all of us were to take a few minutes
and to think about these things, we could all
[16:30] come up with five or 10 of these types of things
very quickly. Whether it's something we do on our
[16:36] own Monday mornings as we start a new week
or something that our team does every week,
[16:40] there are many repeatable tasks. These should
be very very high on your priority list. This is
[16:46] the lowest of hanging fruit in any organization.
It's not sexy, but the back op stuff is very easy,
[16:53] very high ROI. And then the last piece that we've
heard a lot of questions about this week is people
[16:59] are asking, "How do I get a return on investment?
How do I measure return? How do I know if I'm
[17:05] doing this and getting something out of it? And
the reality of that answer is that's really hard.
[17:11] Now, some things are very easy to measure. If I
have an input and an output that equals dollars,
[17:18] that's very easy easy to measure. If I have if
I've added intelligence into something where
[17:24] people are involved and I can see that it takes
20% less time, well, I've just got 20% of an FTE.
[17:31] But not all of the math is that simple. So where
we we encourage a lot of our clients to look at
[17:37] is thinking about return at least initially as
maturity. Find some maturity assessment. There's a
[17:45] lot of them out there. We have one. I'll show you
a QR code at the end. You're welcome to use ours.
[17:50] But get yourself a baseline. Go back today, this
week, tomorrow, and take the test and say, "Okay,
[17:56] as an organization, we scored this." Then in three
or six months from now, go back and take that test
[18:03] again. And you want to see growth in that number.
But when the executive team or the board comes
[18:09] and asks you, hey, what's the return on our
investment? Now all of a sudden, you have this
[18:14] tangible thing you can point to and say, well, in
in August of 26, we took this test and we scored
[18:21] a 2.67, and now in February of 27, we are 3.1.
So the investment we made last year has allowed
[18:30] us to increase maturity by6. That's a lot more
than most organizations can say right now. Most
[18:37] organizations cannot give you a number associated
with ROI. So that is the lowhanging fruit of ROI.
[18:46] Now if you take this one thing back and you have
this one thing for your executive team, I think
[18:51] this is value enough for for the week. But we're
going to spend some time over here on the side at
[18:56] the end of this this uh this session. If you have
questions, I'm happy to answer them. We also are
[19:02] in in the the exhibit hall at booth 1125. We'll be
happy to have a conversation there as well. But I
[19:08] appreciate all of you for your time and and look
forward to having conversations with you later.
