SkillUp Replaces AI Course Completion With Real-Task Capability Audits

Moving Beyond Completion to Capability: Proving AI Upskilling Actually Works — Ai4

Ai4

Enterprise AI investments fail to generate ROI when measured by course completion rather than durable worker capability verified through daily workflow execution.

AI-assisted summary of Moving Beyond Completion to Capability: Proving AI Upskilling Actually Works — original source. Check the original for full context.

Source published: 2026-08-15

Detailed notes

## Background Published on YouTube by Ai4 on August 15, 2026, this session features SkillUp Group executives JPS Kohli, Ratan Deep Singh, and Benjamin George. The presentation addresses the enterprise divide between paying for advanced AI tools with high course completion rates and the lack of real on-the-job capability, presenting a live measurement model backed by human evaluation to demonstrate measurable workplace productivity gains.

## TL;DR The speakers challenge enterprise training reliance on passive course completions, arguing through a bicycle analogy that true capability requires unassisted, repeated performance on real-world deliverables. SkillUp outlines its four-stage framework—Assess, Upskill, Activate, and Transform—demonstrating through a financial analyst persona how the 'Capability Studio' combines an AI Project Companion with human mentor review. The talk emphasizes that durable enterprise capability requires marrying hard technical prompting skills with critical human judgment, safety compliance, and expert coaching to achieve verifiable business value.

## Atmosphere & style The presentation opens with an engaging and direct audience diagnostic that quickly shifts the room into a candid, reflective mood about the shortcomings of corporate training. The progression across the three speakers feels collaborative, brisk, and pragmatic, moving seamlessly from relatable personal storytelling to structured platform demonstrations and pedagogical frameworks. Overall, the talk carries a confident, solution-oriented energy that blends executive urgency with an optimistic, human-centric philosophy.

## Key takeaways - [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) LMS completion metrics track passive activity rather than genuine business execution, creating a misleading perception of workforce AI readiness. - [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) Building AI capability mirrors learning to ride a bicycle: true proficiency only begins when learners can perform unassisted across difficult, real-world conditions. - [03:11](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=191s) Durable capability cannot be evaluated through quizzes or synthetic sandboxes; it must be forged using real daily workplace deliverables. - [03:11](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=191s) The SkillUp framework structures workforce upskilling into four distinct stages: Assess, Upskill, Activate, and Transform. - [03:11](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=191s) Tracking capability metrics over activity provides transparent performance telemetry simultaneously to learners, direct managers, and executive leadership. - [06:35](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=395s) Scaffolded AI Project Companions guide learning by prompting deeper analysis and identifying logical flaws rather than doing the work for the user. - [06:35](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=395s) Effective capability validation combines real-time AI feedback with human mentor review to evaluate process, reasoning, and final artifacts. - [10:37](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=637s) Technical AI skills like prompting and agent orchestration are insufficient without durable human capabilities such as critical judgment, analytical reasoning, and safety verification. - [10:37](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=637s) A 'Human-led, AI-powered' operating model leverages AI for scalable scaffolding while relying on human coaches for orchestration and nuanced strategic guidance. - [14:39](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=879s) Demonstrated workforce AI capability, rather than passive content consumption or certification volume, will be the primary enterprise competitive advantage in the coming decade.

## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) Completion vs. Capability: The Bicycle Analogy

JPS Kohli opens the session by engaging the audience with a diagnostic thought experiment. He asks the room to raise their hands if they have completed an AI course—a prompt that puts hands in the air across the audience. He then delivers the real test: "Keep your hands up if you use everything you've learned every single day." As hands quickly drop, Kohli pinpoints the exact crisis in enterprise learning: organizations have plenty of data showing who completed training, but virtually none showing who can actually perform the work. While course completion merely tracks passive activity like watching videos and clicking through modules, true capability measures performance and business execution.

To illustrate the gap between initial exposure and true capability, Kohli shares a personal story about teaching his daughter how to ride a bicycle. After holding onto the seat and letting go, she managed to balance and pedal for ten seconds. While both father and daughter were ecstatic in the moment, she was far from a capable cyclist. Genuine capability only arrived months later, following scraped knees, training wheels, and relentless practice. She became capable when she could ride uphill, downhill, and across dirt trails completely unassisted, without her father or a coach hovering over her.

Kohli emphasizes that AI learning follows the exact same trajectory. Foundational knowledge, hands-on application, and self-confidence are all necessary stepping stones, but genuine capability only begins when a learner can repeatedly perform tasks without supervision. Transformation occurs when that demonstrated capability translates directly into measurable business value aligned with organizational objectives.

```mermaid flowchart LR accTitle: Progression from Knowledge to Business Transformation subgraph Foundation["Learning Stage"] A["Knowledge & Theory"] --> B["Initial Application"] B --> C["Confidence Building"] end subgraph Mastery["Execution Stage"] C --> D["Repeated Performance"] D --> E["Independent Capability"] E --> F["Business Transformation & ROI"] end ```

[03:11](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=191s) The SkillUp Framework: Shifting from Activity to Real Work

Moving from theory to methodology, Kohli explains that capability cannot be built or proven through another multiple-choice quiz, a generic certificate, or an artificial simulation. Instead, it must be forged using the real work sitting in an employee's daily workflow—such as drafting client proposals, building marketing campaigns, conducting financial analyses, or updating corporate policy documents. By applying AI tools to real-world tasks, receiving expert feedback, and iterating through a rigorous "review, improve, repeat, rinse" cycle, surface-level confidence matures into durable capability.

While SkillUp’s full enterprise framework spans four distinct stages—Assess, Upskill, Activate, and Transform—Kohli highlights the two pillars that actively change on-the-job behavior: Activate and Transform. In the Activate stage, learners work on authentic projects inside a specialized environment where both AI systems and human mentors provide direct feedback on their execution. Rather than evaluating whether an employee finished a syllabus, the evaluation focuses on prompt quality, output quality, receptiveness to critique, and iterative refinement until the worker can solve problems independently.

Once an employee demonstrates autonomy, they advance to the Transform phase, where they are tasked with solving high-impact enterprise problems. This structured approach makes organizational capability completely transparent. Learners see their own technical progression, managers gain clear visibility into where team members excel or struggle on real deliverables, and corporate leaders can pinpoint exact workforce proficiencies rather than relying on assumptions.

- **Activity Metrics vs. Capability Metrics**: Shift from tracking video watch time and module completions to evaluating prompt engineering depth, output accuracy, feedback iteration, and unassisted execution. - **Workflow-Embedded Practice**: Projects utilize real workplace artifacts (financial models, proposals, marketing collateral) rather than synthetic sandbox exercises. - **Three-Tier Visibility**: Performance telemetry is exposed simultaneously to individual learners, direct managers, and executive leadership to guide strategic reskilling.

[06:35](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=395s) Demonstration: Maya’s Workflow in the Capability Studio

Benjamin George takes over to showcase how this philosophy operates within SkillUp's software environment, walking through the journey of a persona named Maya, a Senior Financial Analyst at Northstar Financial Group. Maya's goal is to streamline and modernize her daily analytical workflows. Having already completed SkillUp’s progressive assessment—which continuously updates as an employee advances—and her core upskilling cohort, Maya enters the platform’s "Capability Studio."

Inside the Capability Studio, Maya engages with AI-adaptive simulations designed around real-world workplace scenarios. These simulations do not merely test technical recall; they actively cultivate durable judgment and adaptability. Upon completing a simulation, Maya submits her work alongside written reflections and supporting evidence. The platform immediately generates an automated AI evaluation score, which is then routed directly to SkillUp human mentors for validation and in-depth review.

```mermaid flowchart TD accTitle: Capability Studio Project Workflow subgraph Studio["Capability Studio Execution"] A["Define Purpose, Tasks & Success Metrics"] --> B["Iterate with AI Project Companion"] B --> C["Submit Evidence & Final Artifacts"] end subgraph Review["Human-in-the-Loop Validation"] C --> D["Submit Learner Reflections"] D --> E["Instant AI Scoring"] E --> F["Human Mentor Review & Validation"] F --> G["Measured Impact: Hundreds of Hours Saved"] end ```

George demonstrates an active project Maya is navigating. Each assignment clearly outlines its purpose, step-by-step tasks, and specific success metrics. To guide her, Maya collaborates with an "AI Project Companion"—a scaffolded assistant that intentionally avoids doing the work for her. Instead, the companion identifies logical flaws in her reasoning, prompts deeper analysis, and pushes her through iterative cycles where breakthrough "aha moments" take place. After uploading her final results, proof of execution, and self-reflections on what she learned and where she can improve, the submission is reviewed by human experts, confirming that her streamlined workflow saves hundreds of hours per annum.

[10:37](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=637s) The Secret Sauce: Human-Led AI Orchestration and Durable Skills

Ratan Deep Singh takes the stage to unpack the key architectural and pedagogical lessons from the demonstration. He reiterates the four-stage framework—Assess, Upskill, Activate, and Transform—noting that Maya began with diagnostic assessments to uncover her baseline skills relative to organizational expectations, progressed through customized, scaffolded pedagogy, and then applied those skills to hands-on project workflows in Activate.

Singh highlights the critical balance of technology and mentorship: while Maya interacted with specialized AI models and agents, the true "secret sauce" behind her success was SkillUp's human instructors, coaches, and mentors. These experts worked directly alongside her to provide orchestration, nuance, and strategic guidance, embodying SkillUp's foundational operating model: *Human-led, AI-powered*.

```mermaid flowchart LR accTitle: Balanced AI Skill Architecture subgraph HardSkills["Technical AI Skills"] A["Tool Proficiency"] B["Prompt Engineering"] C["Agentic Workflows"] end subgraph DurableSkills["Durable Capabilities"] D["Critical Judgment"] E["Analytical Reasoning"] F["Safety & Verification Mindset"] end HardSkills --> G["Comprehensive Workforce Capability"] DurableSkills --> G ```

Furthermore, Singh critiques typical market offerings that focus solely on hard technical skills—such as tool syntax, basic prompt writing, or assembling agent workflows. Hard skills alone are insufficient without durable human capabilities: critical thinking, sound business judgment, cross-functional collaboration, and an adaptive, safety-conscious mindset. Throughout her project cycles, Maya was forced to interrogate AI outputs, refine her judgment, and verify results against enterprise AI safety standards.

Singh notes that SkillUp has delivered AI upskilling for over a decade, but enterprise demand has surged dramatically. The organization has expanded its active global AI readiness deployments across enterprises, academia, and government from 25+ to 32 programs within a two-day window, delivering documented ROI and productivity gains.

- **Human-in-the-Loop Pedagogy**: Advanced AI agents handle real-time feedback and evaluation scaffolding, while human master coaches oversee skill verification and orchestration. - **Holistic Skill Development**: Combines technical hard skills (prompting, agent architectures) with durable capabilities (critical thinking, output validation, compliance with enterprise safety guidelines). - **Global Adoption**: Rapid expansion from 25+ to 32 live enterprise, academic, and governmental AI readiness programs worldwide.

[14:39](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=879s) The Decade Ahead: Capability as a Competitive Advantage

JPS Kohli returns for closing remarks, circling back to his opening question. He challenges the audience to look past internal training completion metrics and confront the harder reality: "How many people in your organization can actually demonstrate AI capability?" Over the coming decade, this distinction will cease to be a mere human resources or training metric; it will represent an enterprise's primary competitive advantage.

Kohli concludes with SkillUp's core philosophy: human potential remains the world's greatest resource, and AI serves to amplify it rather than replace it. Lasting enterprise value is never generated by pushing more passive content onto employees, but by cultivating compounding, verifiable capabilities that drive tangible business outcomes.

## Notable quotes > Your company already pays for Copilot, maybe Claude too, and most of your people use them like expensive autocomplete.

> Meanwhile the LMS swears everyone's trained. This talk is about the gap between those two facts and what it costs you every quarter.

> We'll run our measurement model live on stage: real work scored by human reviewers, and dashboards that show whether people actually got faster and better, not whether they clicked Next until a certificate appeared.

> Come skeptical. We're bringing receipts.

> [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) Keep your hands up if you use everything you've learned every single day. — JPS Kohli

> [03:11](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=191s) review, improve, repeat, rinse — JPS Kohli

> [10:37](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=637s) Human-led, AI-powered — Ratan Deep Singh

> [14:39](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=879s) How many people in your organization can actually demonstrate AI capability? — JPS Kohli

## Mentioned Ai4, JPS Kohli, Ratan Deep Singh, Benjamin George, SkillUp Group, SkillUp Online, Copilot, Claude, Northstar Financial Group, Capability Studio, AI Project Companion

## Video description JPS Kohli - Founder and Group CEO, SkillUp Group Ratan Deep Singh - CEO, SkillUp Online Benjamin George - Global AI Leader, SkillUp Group

Your company already pays for Copilot, maybe Claude too, and most of your people use them like expensive autocomplete. Meanwhile the LMS swears everyone's trained. This talk is about the gap between those two facts and what it costs you every quarter. We'll run our measurement model live on stage: real work scored by human reviewers, and dashboards that show whether people actually got faster and better, not whether they clicked Next until a certificate appeared. We'll walk through how one enterprise turned a floor of casual users into people who measurably work differently. Come skeptical. We're bringing receipts.

[Source](https://www.youtube.com/watch?v=6Cbbuut4vI0)

中文简报

SkillUp以真实任务能力审计取代AI课程完课率如果仅以课程完成度而非通过日常工作流执行所验证的员工持久能力来衡量，企业AI投资将难以产生投资回报率（ROI）。演讲者质疑了企业培训对被动完课率的依赖，并通过骑自行车的类比指出，真正的能力需要员工在没有辅助的情况下，在实际工作交付物上反复实践来体现。SkillUp阐述了其包含“评估（Assess）、技能提升（Upskill）、激活（Activate）与转型（Transform）”的四阶段框架，并通过财务分析师的角色画像，展示了“Capability Studio”如何将AI项目伴随助手与人工导师评审相结合。本次演讲强调，持久的企业能力需要将扎实的技术提示词技能与批判性的人类判断、安全合规以及专家辅导相结合，以实现可验证的商业价值。 LMS完课指标追踪的是被动活动而非真实的业务执行，从而对员工队伍的AI就绪度产生了误导性认知。 (0:04) 培养AI能力就像学骑自行车：只有当学习者能够在复杂的现实条件下无需协助独立操作时，真正的熟练才算开始。 (0:04) 持久的能力无法通过测验或模拟沙盒来评估；它必须借助日常真实的职场交付物来锤炼。 (3:11) SkillUp框架将员工技能提升结构化为四个不同阶段：评估（Assess）、技能提升（Upskill）、激活（Activate）和转型（Transform）。 (3:11) 跟踪能力指标而非活动指标，能同时向学习者、直接主管和高管层提供透明的绩效遥测数据。 (3:11)

Reading summary

JPS Kohli and the SkillUp team argue that enterprise AI training fails when organizations measure passive course completion rather than independent execution on authentic workplace tasks. Achieving measurable business value requires shifting to workflow-embedded practice that combines technical fluency, durable cognitive judgment, and human-in-the-loop mentorship. Completion Metrics vs. Autonomous Capability (0:04) Kohli asserts that while corporate workers frequently finish introductory AI courses, very few apply what they learn to their daily routines. Tracking video watch time and multiple-choice quiz completions captures passive exposure rather than actual execution. Comparing the learning curve to a child mastering a bicycle, Kohli notes that initial unassisted balance is merely a starting point; genuine capability requires repeated, unassisted performance across difficult conditions such as hills and dirt trails. In an enterprise setting, genuine capability begins only when an employee can execute workflows repeatedly without supervisory intervention. Kohli emphasizes that business transformation and return on investment occur when this demonstrated autonomy translates directly into organizational priorities. Grounding Upskilling in Daily Workplace Workflows (3:11) To build durable skills, SkillUp argues that training must abandon synthetic sandbox exercises and focus on real workplace deliverables, such as drafting proposals, financial models, marketing campaigns, or corporate policies. Within its four-stage framework—Assess, Upskill, Activate, and Transform—the Activate and Transform stages actively reshape job behavior. Activate immerses employees in real projects where their prompt depth, output quality, and iterative refinement are evaluated, while Transform tasks autonomous employees with solving high-impact enterprise problems. This workflow-embedded approach introduces three-tier telemetry. Performance data is shared simultaneously with individual learners, direct managers, and executive leadership, replacing corporate assumptions with transparent visibility into specific workforce proficiencies. Scaffolding Guidance Through Human and AI Mentorship (6:35) Demonstrating the platform through Maya, a senior financial analyst persona, Benjamin George shows how SkillUp's "Capability Studio" structures daily analytical tasks. Maya interacts with an "AI Project Companion" programmed not to perform tasks for her, but to question her reasoning, highlight logical gaps, and guide her through iterative refinement. Once a project is submitted alongside learner reflections, automated AI scoring is paired with evaluation from human mentors. Ratan Deep Singh explains that this "human-led, AI-powered" model is essential; while AI agents handle real-time scaffolding, human coaches provide strategic nuance, verify project artifacts, and validate reported efficiencies, such as the hundreds of annual hours saved in Maya's analytical workflow. Integrating Durable Cognitive Skills with Technical Fluency (10:37) Singh critiques market training programs that focus exclusively on hard technical mechanics, such as prompt syntax or configuring agentic pipelines. Technical fluency remains insufficient without durable human capabilities, including critical reasoning, business judgment, cross-functional collaboration, and compliance with enterprise AI safety standards. Throughout her project cycles, Maya was required to interrogate model outputs and verify data rather than passively accept AI-generated answers. Singh notes that organizational demand for this balanced pedagogy has surged, reporting that SkillUp expanded its active global AI readiness programs across corporate, academic, and government clients from 25-plus to 32 within a two-day window. Kohli forecasts that over the coming decade, an organization's primary competitive advantage will depend on verifiable workforce capability rather than nominal training completion. The central challenge for enterprise leadership remains moving away from passive content consumption to establish continuous, human-validated execution across operational workflows.

精炼阅读摘要

SkillUp代表在Ai4大会上提出，企业AI培训已面临严重的有效性危机：传统的课程完成率仅反映员工的被动参与，无法证明实际工作能力。唯有通过嵌入日常业务场景、结合AI即时反馈与人类导师核验的实践，企业才能将技能转化为经得起检验的商业价值与竞争壁益。完成率与真实能力的脱节 (0:04) JPS Kohli在演讲开场指出，许多企业虽然拥有大量员工完成AI培训课程的数据，但在实际业务中能够每天自主运用所学技能的人却寥寥无几。课程完成率通常只记录了观看视频和点击模块等被动行为，既不能反映员工的业务执行力，也无法衡量其实际工作表现。 Kohli以教女儿骑自行车的经历作比喻：初次能短时间平衡并不代表具备独立骑行能力，真正的能力需要经历持续练习、摔跤和拆除辅助轮，最终能在无人看护的情况下应对上下坡及复杂泥地。AI技能的学习同样遵循此规律，从基础理论、初步应用到建立自信只是铺垫，真正的能力始于员工无需外界监督即可反复独立完成任务，并为企业创造出切实的业务价值。基于真实工作流的培养框架 (3:11) 针对传统考核方式的不足，Kohli指出多选题测验、通用证书或虚构模拟无法构建并证明真正的专业能力。持久的能力必须在员工日常接触的实际工作中锻造，例如撰写客户提案、制定营销方案、开展财务分析或更新企业制度文件。通过将AI工具应用于真实产出，并经过专家反馈及“审查、改进、重复”的迭代循环，员工的技能才能真正沉淀。 SkillUp构建了涵盖“评估（Assess）、提升（Upskill）、激活（Activate）、转型（Transform）”的四阶段企业框架。其中Activate阶段让学员在专用环境中处理真实项目，由AI系统与人类导师协同提供针对Prompt质量、输出水准和迭代修改的直接反馈；当员工展现出自主工作能力后，便进入Transform阶段去解决高价值的企业实际难题。该模式同时向员工、直属主管及高管呈现能力数据，摒弃了单纯跟踪视频时长等活动指标的做法。能力工作坊中的工作流验证 (6:35) Benjamin George展示了该方案在软件系统中的具体运作，以虚构角色Northstar Financial Group的高级财务分析师Maya为例。Maya在完成渐进式能力评估和培训阶段后，进入“能力工作坊（Capability Studio）”，参与旨在培养长效判断力与适应力的真实业务模拟工作，其目标在于优化并更新现有的日常分析流程。在演示的项目任务中，系统明确列出工作目标、执行步骤与成功评估指标。Maya通过与系统内的“AI项目助手（AI Project Companion）”进行协作，该助手不会直接代劳任务，而是指出其思考中的逻辑缺陷、引导深入推导并促成认知突破。完成方案后，Maya提交最终产出、执行证据与自我反思，系统会立即生成AI评分，随后转交SkillUp的人类导师进行人工复核与验证。据George说明，此类流程优化能够为企业年均节省数百小时的工作量。人机协同与持久素养的结合 (10:37) Ratan Deep Singh对技术演示进行了教学逻辑拆解，强调该模式的核心运行机制为“以人为本，AI驱动（Human-led, AI-powered）”。尽管学员在日常流程中频繁与专业AI模型和智能代理交互，但 SkillUp 的人类讲师、教练与导师在全程提供策略指引、细致研判与最终验证，构成了不可或缺的关键环节。此外，Singh批评了当前市场上仅侧重工具语法、基础Prompt编写或智能代理编排等硬技能的培训路径。他主张技术硬技能必须与批判性思维、业务判断力、跨部门协作以及符合企业安全准则的核验意识等“持久能力（durable skills）”相结合，学员必须学会在实际业务中审视AI生成的内容并验证结果。Singh提到，SkillUp在该领域的全球AI就绪部署项目已在短时间内从25个以上扩展至32个，覆盖企业、高校与政府机构。 Kohli在总结中预测，在未来十年中，衡量拥有经得起检验的AI能力的员工规模将超越单纯的HR考评范畴，演变为企业的核心竞争壁垒。不过，该演讲主要依赖于SkillUp自身的方法论框架与产品演示，至于各类企业在面对多样化业务时，实际推行真实工作流项目培训与人类导师核验机制的落地成本与长期转化效率，仍有待更广泛的外部验证。

Detailed notes — zh

## Background 该视频于 2026 年 8 月 15 日由 Ai4 发布在 YouTube 平台上。内容由 SkillUp Group 核心管理层（JPS Kohli、Ratan Deep Singh 与 Benjamin George）共同主讲，旨在揭示企业在 AI 培训中盲目追求 LMS“课程完成度”而非“实际工作能力”的普遍痛点，并现场展示如何通过真实业务场景与评估模型量化 AI 技能培训的投资回报率。 ## TL;DR 演讲直击传统 AI 培训浮于表面的弊端，提出衡量员工 AI 水平的核心标准在于脱离指导后稳定交付业务成果的能力。SkillUp 团队拆解了“评估、提升、激活、转化”的四步进阶赋能框架，并通过金融分析师 Maya 的真实成长案例，生动演示了如何借助“能力工作室”中的苏格拉底式 AI 伴侣与人类导师双轨审核闭环大幅节省工时。最终，演讲强调真正的 AI 胜任力必须将提示词等硬技能与批判性思维、商业判断等长青技能深度结合，以此构建企业未来十年的核心竞争壁垒。 ## Atmosphere & style 整场演讲节奏紧凑且极具实战说服力，三位演讲者交替登台、配合默契，开场直击痛点的互动与生动的骑车比喻迅速拉近了与观众的距离。分享摒弃了空洞的行业套话，从对形式主义培训的犀利剖析，自然过渡到严谨细致的平台实操与数据佐证，全程充满笃定与理性。收尾处的升华既饱含对人类潜能的温度与敬意，又带给在场企业决策者一种紧迫而清晰的行动方向感。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) 传统的课程“完成度”（Completion）仅代表员工刷视频、点击下一步直至拿证的活动量，无法衡量员工在实际工作中独立交付成果的真实“能力”（Capability）。 - [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) AI 技能的习得如同学习骑自行车，单次尝试或初步建立信心并不等于掌握能力，真正的能力体现在脱离指导后能够一次又一次稳定交付业务成果。 - [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) AI 能力构建不能依赖脱离实际的随堂测试或纸上证书，必须依托员工日常真实的业务载体（如提案、营销方案、财务报告等）进行实战演练。 - [04:00](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=240s) SkillUp 提出了从“评估（Assess）、提升（Upskill）、激活（Activate）到转化（Transform）”的四步进阶框架，核心在于通过真实项目与双轨反馈促成行为改变。 - [04:00](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=240s) 新型评估体系深入考察提示词质量、模型输出水平、对导师反馈的消化改进程度，以及零干预独立交付能力，使组织各层级的能力进阶完全可视化。 - [06:35](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=395s) “能力工作室”（Capability Studio）结合 AI 自适应模拟与“AI 项目伴侣”，以苏格拉底式提问引导学员发现思维盲区，而非直接提供现成答案。 - [06:35](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=395s) SkillUp 采用“AI 驱动、人类引领”（AI-powered, Human-led）的双轨审核机制，结合 AI 实时初审与人类导师深度复核，确保学习闭环与业务产出质量。 - [10:36](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=636s) 市面上许多 AI 培训局限于工具调用和提示词编写等“硬技能”，而企业员工同样需要打磨商业研判力、批判性思维与安全合规意识等“长青技能”（Durable Skills）。 - [14:39](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=879s) 在接下来的十年中，企业中能够真正“展示并产出”AI 能力的员工规模，将从单纯的人力资源统计数据演变为决定企业生死存亡的核心竞争壁垒。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) 从“课程完成”到“实际能力”：重新定义 AI 培训成效演讲伊始，SkillUp Group 创始人兼集团首席执行官 JPS Kohli 向在场观众抛出了两个直击痛点的问题：“如果你已经完成了一门 AI 课程，请举手；如果你每天都在实际工作中运用所学的内容，请继续举着手。”现场绝大多数举起的手瞬间放了下来。JPS 直言：“这就是我们今天站在这里的原因。”如今每家企业都能拿出漂亮的数据，精确统计有多少员工“完成了培训”或“刷完了视频”，但极少有企业能回答那个更为严峻的问题——员工是否真正具备了完成工作的实际能力？“完成度”（Completion）衡量的仅仅是活动量，而“能力”（Capability）衡量的则是真实的绩效产出。 ```mermaid flowchart LR accTitle: Learning to Business Value Flow subgraph Foundation ["基础阶段"] K["知识积累 (Knowledge)"] --> A["实际应用 (Application)"] A --> C["建立信心 (Confidence)"] end subgraph Mastery ["能力沉淀"] C --> R["重复演练 (Repeated Performance)"] R --> CAP["真正能力 (Capability)"] end subgraph Impact ["业务赋能"] CAP --> B["业务价值与组织对齐 (Business Value)"] end ``` 为了阐明这一质的飞跃，JPS 分享了自己教女儿骑自行车的亲身经历。当他第一次小心翼翼地松开扶着车座的手，女儿摇摇晃晃地骑了十秒钟，两人当时都兴奋无比；但这并不意味着女儿在当天就掌握了骑车这项能力。真正形成能力是在数月之后——在经历了无数次摔倒、重新站起、借助辅助轮以及反反复复的练习后，女儿不仅能应对上坡、下坡和泥泞泥地，更重要的是，她完全不再需要父亲或教练在旁指点，能够独立自如地骑行。AI 技能的习得同样遵循完全相同的规律：真正的能力并非诞生于获取知识、单次应用或初步建立信心的瞬间，而是始于在脱离指导的情况下能够“一次又一次重复稳定地完成工作”，并最终将这种能力转化为与企业目标高度一致的商业价值。 - **活动与绩效的本质区别**：完成度仅代表员工点击“下一步”直到获取证书的被动行为；能力则体现为独立、稳定输出业务成果的绩效表现。 - **能力构建的真实载体**：提升能力无法依赖更多的随堂测试、模拟练习或纸上证书，必须依托员工收件箱中真实的日常业务——撰写提案、制定营销方案、编制财务分析报告或拟定政策文件。 - **循环迭代机制**：在真实业务场景中利用 AI 辅助与人类导师的介入，通过“审阅、改进、重复、沉淀”的闭环，促使操作信心真正蜕变为主体能力。 --- ### [04:00](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=240s) SkillUp 四步进阶框架：激活与转化的核心支柱 JPS 介绍了 SkillUp 经过实践沉淀出的完整 AI 赋能框架。虽然整个框架包含多个阶段，但在有限的时间内，他重点拆解了能够真正促成行为改变的两大支柱环节：“激活”（Activate）与“转化”（Transform）。在激活阶段，SkillUp 在当年搭建了专门的实操平台，员工不再纸上谈兵，而是在真实项目中进行高频演练，并实时接收来自 AI 和人类导师的双轨反馈。通过吸收反馈并不断修正，学员能够深刻感知并提升自身的实际绩效；一旦达到这一标准，他们便能顺畅跨入转化阶段，独立攻克企业面临的真实业务挑战。 ```mermaid flowchart TD accTitle: SkillUp Four-Step Progression Framework subgraph Pipeline ["SkillUp 赋能闭环"] P1["1. 评估 (Assess) 精准诊断技能断层"] --> P2["2. 提升 (Upskill) 搭建知识与技能支架"] P2 --> P3["3. 激活 (Activate) 真实项目演练与双轨反馈"] P3 --> P4["4. 转化 (Transform) 独立解决复杂业务问题"] end ``` 这种模式彻底重塑了评估维度。评估重点不再停留在“是否观看了视频”或“是否完成了某个模块”，而是深入考察提示词工程的质量（Prompt Quality）、模型输出的实际质量、学员对导师反馈的消化与改进程度，以及随着时间推移能否在零干预下独立交付。这种机制让能力变得清晰“可见”：学员能够直观看见自己的能力进阶轨迹；团队主管能明确掌握下属在哪些项目中表现优异、在何处遭遇瓶颈；企业管理层则可以依据详实的数据资产，清晰洞察组织在哪些业务板块已构建起坚实的 AI 能力，而在哪些领域仍需加大投入。 --- ### [06:35](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=395s) 平台实操演示：金融分析师 Maya 的能力跃迁之旅随后，SkillUp Group 全球 AI 负责人 Benjamin George 接过演示，通过展示平台上一名真实学员的成长旅程，将抽象的方法论具象化。演示主角是来自 Northstar Financial Group 的高级财务分析师 Maya，她的核心诉求是利用 AI 重构日常分析流程以大幅提升工作效能。在进入核心训练前，Maya 已经完成了动态进阶评估（随着学习进程不断刷新技能画像）以及专属的定制化知识提升阶段，随后正式解锁了平台的核心模块——“能力工作室”（Capability Studio）。在能力工作室中，Maya 首先经历了一系列“AI 自适应模拟场景”。这些模拟紧密贴合日常高频金融业务，旨在打磨其关键的批判性思维与判断力等长青技能。模拟结束后，Maya 提交了自己的分析依据与反思总结。系统不仅由 AI 在后台即刻生成多维度的初步评估得分，还会将完整的交互证据链打包分发给 SkillUp 的专业人类导师进行二次深度审核。 ```mermaid flowchart TD accTitle: Capability Studio Workflow subgraph Workflow ["能力工作室实战工作流"] S["自适应模拟与项目启动"] --> C["AI 项目伴侣协作 (启迪逻辑/暴露盲区)"] C --> E["提交工作证据与反思日志"] E --> A["AI 实时多维初审评分"] A --> H["人类导师专业复核与精准反馈"] H --> R["沉淀为可量化的企业业务价值"] end ``` 在随后的实际项目演练中，Maya 明确知晓项目的业务目标、具体任务与成功指标。在此过程中，“AI 项目伴侣”（AI Project Companion）全程协同——它从不直接替学员生成最终答案，而是通过层层提问指引其思维盲区，激发其自主迭代与推演。完成所有分析后，Maya 附上完整的佐证数据与反思日志（记录心得、不足与迭代方向）并提交审核。这种“AI 驱动、人类引领”（AI-powered, Human-led）的机制为 Maya 带来了显著的蜕变：数据量化显示，她每年在常规工作流中节省了数百小时的工时，实现了实打实的业务价值转化。 - **AI 项目伴侣定位**：扮演“苏格拉底式”引导者，指出逻辑漏洞与推演死角，促成学员产生自主领悟的“顿悟时刻”（Aha moment）。 - **实证交付物闭环**：要求提交详尽的操作佐证（Evidence）与反思记录（Reflections），杜绝虚假完成。 - **量化业务产出**：通过真实工作流重塑，单人每年可节约数百小时工时，让 ROI（投资回报率）清晰可见。 --- ### [10:36](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=636s) 成功秘诀：人类导师引领与硬技能之外的“长青技能” SkillUp Online 首席执行官 Ratan Deep Singh 随后登台总结了演示背后的深层逻辑。他强调，Maya 之所以能成功完成从评估、提升、激活到转化的四步跨越，固然离不开 SkillUp 先进的 AI 模型与智能体系统的支撑，但真正的“秘密武器”在于 SkillUp 背后的人类导师团队。正是这些经验丰富的讲师、导师和教练全程紧密陪伴，在关键节点提供方向指引，并进行精细的 AI 流程编排，才确保了学员能以最高效的方式构建能力。这完美契合了 SkillUp 的核心哲学——“以人为本，AI 赋能”（Human-led, AI-powered）。 ```mermaid flowchart TD accTitle: Dual Skill Dimensions for AI Mastery subgraph Skills ["全面 AI 能力模型"] subgraph Hard ["硬核 AI 技能 (Hard Skills)"] H1["工具与平台操作"] H2["提示词工程 (Prompting)"] H3["智能体工作流设计 (Agentic Workflows)"] end subgraph Durable ["长青长效技能 (Durable Skills)"] D1["专业商业研判力 (Judgment)"] D2["批判性思维 (Critical Thinking)"] D3["安全合规验证与伦理把控 (Safety & Standards)"] end end ``` Ratan 进一步指出，当前市面上绝大部分 AI 培训往往存在严重的结构性短视——课程开始于也终结于纯粹的“硬核 AI 技能”（如编写提示词、调用 AI 工具或搭建 Agentic 工作流）。这些技能固然重要，但远非全部。企业员工更需要具备“长青技能”（Durable Skills）与适应性思维，才能在快速变化的环境中做出审慎的商业决策。在整个激活训练期间，Maya 的商业判断力随每一次项目迭代而提升，批判性思维在与 AI 的多轮辩论中被持续激活，同时其验证与审核能力也大幅增强，确保所有输出严格符合企业既定的 AI 安全与合规准则。 - **项目规模迅速扩张**：SkillUp 在企业、政府及学术界的 AI 赋能领域深耕已逾十年；自推出该“AI 战备计划”（AI Readiness Program）以来，全球推进的落地项目已在短短几天内从 25 个迅速攀升至 32 个。 - **复合型能力诉求**：真正的 AI 胜任力是“硬技能”（操作与工作流编排）与“长青技能”（判断力、批判性思维与安全合规意识）的深度结合。 --- ### [14:39](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=879s) 结语：将 AI 能力转化为未来十年的核心竞争优势在大会尾声，JPS Kohli 重新登台作总结发言。他呼应开场的提问，向在座所有企业决策者提出了一个更为深远的问题：“在你的企业中，究竟有多少人能够真正‘展示’出 AI 能力？”他郑重指出，在接下来的十年中，这一指标将不再仅仅是企业人力资源部门的一项培训统计数据，它将直接演变为决定企业生死存亡的核心竞争壁垒。 JPS 动情地总结道，SkillUp 始终坚信人类潜能依然是这个世界上最伟大的资源，而 AI 的本质是放大人类潜能的强力杠杆。企业的持久价值绝非通过向员工灌输更多泛泛的课程所能创造，唯有依靠在实战中沉淀那些能够产生复利效应的真本领，并将其深度转化为切实的业务价值与商业影响力，才能在智能化浪潮中立于不败之地。 ## Notable quotes > [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) 如果你已经完成了一门 AI 课程，请举手；如果你每天都在实际工作中运用所学的内容，请继续举着手。 — JPS Kohli > [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) 这就是我们今天站在这里的原因。 — JPS Kohli > [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) 完成度衡量的仅仅是活动量，而能力衡量的则是真实的绩效产出。 — JPS Kohli > [00:04](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=4s) 一次又一次重复稳定地完成工作 — JPS Kohli > [06:35](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=395s) AI 驱动、人类引领 — Benjamin George > [10:36](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=636s) 以人为本，AI 赋能 — Ratan Deep Singh > [14:39](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=879s) 在你的企业中，究竟有多少人能够真正‘展示’出 AI 能力？ — JPS Kohli > [14:39](https://www.youtube.com/watch?v=6Cbbuut4vI0&t=879s) 人类潜能依然是这个世界上最伟大的资源，而 AI 的本质是放大人类潜能的强力杠杆。 — JPS Kohli ## Mentioned JPS Kohli, Ratan Deep Singh, Benjamin George, SkillUp Group, SkillUp Online, Northstar Financial Group, Maya, Ai4, Copilot, Claude, 能力工作室 (Capability Studio), AI 项目伴侣 (AI Project Companion), AI 战备计划 (AI Readiness Program) ## Video description JPS Kohli - Founder and Group CEO, SkillUp Group Ratan Deep Singh - CEO, SkillUp Online Benjamin George - Global AI Leader, SkillUp Group Your company already pays for Copilot, maybe Claude too, and most of your people use them like expensive autocomplete. Meanwhile the LMS swears everyone's trained. This talk is about the gap between those two facts and what it costs you every quarter. We'll run our measurement model live on stage: real work scored by human reviewers, and dashboards that show whether people actually got faster and better, not whether they clicked Next until a certificate appeared. We'll walk through how one enterprise turned a floor of casual users into people who measurably work differently. Come skeptical. We're bringing receipts. [Source](https://www.youtube.com/watch?v=6Cbbuut4vI0)

Source quotes

[0:04] JPS Kohli: 如果你已经完成了一门 AI 课程，请举手；如果你每天都在实际工作中运用所学的内容，请继续举着手。 [0:04] JPS Kohli: 这就是我们今天站在这里的原因。 [0:04] JPS Kohli: 完成度衡量的仅仅是活动量，而能力衡量的则是真实的绩效产出。 [0:04] JPS Kohli: 一次又一次重复稳定地完成工作 [6:35] Benjamin George: AI 驱动、人类引领 [10:36] Ratan Deep Singh: 以人为本，AI 赋能 [14:39] JPS Kohli: 在你的企业中，究竟有多少人能够真正‘展示’出 AI 能力？ [14:39] JPS Kohli: 人类潜能依然是这个世界上最伟大的资源，而 AI 的本质是放大人类潜能的强力杠杆。

Source quotes

Your company already pays for Copilot, maybe Claude too, and most of your people use them like expensive autocomplete. Meanwhile the LMS swears everyone's trained. This talk is about the gap between those two facts and what it costs you every quarter. We'll run our measurement model live on stage: real work scored by human reviewers, and dashboards that show whether people actually got faster and better, not whether they clicked Next until a certificate appeared. Come skeptical. We're bringing receipts. [0:04] JPS Kohli: Keep your hands up if you use everything you've learned every single day. [3:11] JPS Kohli: review, improve, repeat, rinse [10:37] Ratan Deep Singh: Human-led, AI-powered [14:39] JPS Kohli: How many people in your organization can actually demonstrate AI capability?

Original transcript

[0:04] Let me begin by asking two questions. [0:08] The first one there's hands up if you have completed an AI course. [0:18] Keep your hands up if you use everything you've learned every single day. [0:26] That's why we are here. [0:30] So every organization today can tell you how [snorts] many people have completed the training. [0:37] Very few can answer a much more difficult question which is can they actually do the work. [0:46] Those are completely different measurements. [0:50] Completion measures activity. Did they complete the course? Did they watch the video? Etc. [0:58] Whereas capability measures performance and that's the shift I want you guys to think about today. [1:08] So I'll tell a story. I remember teaching my daughter how to ride a bicycle and I think [1:14] everybody here whoever is a parent would have can easily relate to that story. I held on to the seat [1:22] and then with difficulty I let it go and for 10 seconds she rode the bike. I was thrilled. [1:29] She was ecstatic. But did she become capable a capable rider that day? I don't think so. [1:37] Capability happened months later when she fell when she tried again. I held her hand. She had [1:44] training wheels and after repeated practice she could now ride a bicycle. She could go uphill, [1:50] downhill in the dirt. But more importantly, she was doing it on her own without any instruction [1:57] without her dad or a coach available to guide her. That's when capability happens. And AI learning [2:04] works exactly in the same way. Nothing different. So when does learning actually become capability? [2:14] Is it through knowledge? Well, knowledge is a necessary step. Is it through application? [2:21] Well, that's important too. Is it through building confidence? Well, confidence is necessary, too. [2:28] But the real capability happens or begins when you're able to repeatedly perform [2:35] over and over again. That's when capability becomes real. And then transformation happens [2:41] when you're able to use that capability and create business value. And that business value [2:48] is aligned with your organization goals. And that is real progression that changes how we [2:55] think about learning. And that's the important shift that we at Scaleup also went through. [3:01] We went from meeting completion metrics to really trying to measure outcomes and capability. [3:11] So if capability is the destination, how do you build this capability? Is it through another quiz, [3:19] through another certification, through a simulation? No. It is through real work. [3:27] The work that is already sitting in somebody's inbox. It is a proposal, a marketing campaign, [3:33] a financial analysis or a policy document. Again, real work that is improved through AI [3:40] or through human guidance and nudging. And you review, you improve, you repeat, you rinse [3:47] and over and over again. That's how your confidence truly becomes capability. [3:54] So everything that I've described till now has been under this much larger AI framework [4:01] that Scalop has developed. But given that we only have 20 minutes, I'm going to cheat a bit [4:06] and just cover the two key pillars that actually change behavior. And those two pillars are [4:13] activate and transform. So we we have been able to assess people look at the gaps [4:19] provide them the knowledge and skills um through upskilling but in the activate piece this year [4:27] we built a platform and we've built area where people can actual pro uh practice real projects [4:34] and they get feedback through AI and human mentors and through that feedback process they're able [4:40] to see and perform and that changes the way they really create capability and performance and once [4:48] they're able to do it then we can tell them here is a business problem go and solve that business [4:53] problem and create business value. Hence, if capability is the goal, then how do you evaluate [5:00] the work itself? It's not by evaluating the activity. It's not by did you complete the course, [5:06] did you watch a video or did you complete a module? We look at the quality of the prompt, the [5:12] quality of the output. Did you get feedback? How did you work on that feedback and enhance your own [5:19] work? And you to repeat, learn from it. And over time are you able to do it without any guidance [5:25] or instruction and that's when you can move to a transform phase and solve a real business problem. [5:33] So once the work is assessed capability becomes very visible. Why? Because a learner can now see [5:41] how the learner is improving. A manager can see what is the projects learner performed in [5:47] and where did he fail. and organization can see the data and say well I am actually building [5:53] organization capability in these areas but I am still struggling in these other ones [5:59] and that's where capability actually becomes more visible and [snorts] that's where adoption happens [6:05] it's not something that you're trying to assume it is something that you can actually point to [6:11] and with that note I've actually talked a lot about the philosophy of capability [6:16] but like I said capability is something that's palpable. It's visible. [6:21] I'll now transition over to my colleague Ben who's built this platform. Uh but it's more [6:27] than a demonstration. He'll actually show you how capability sits in action. So Ben, take it away. [6:35] So how do we take philosophy and turn it into capability? Uh [6:40] in order to demonstrate that today, we're going to join one of our cohort members in the AI [6:45] readiness program, Maya. Now Maya is a senior financial analyst at Northstar Financial Group [6:50] and her purpose on the platform is to create more effective and efficient workflows. Now Maya has [6:57] been in our AI readiness journey for a while. So she's already gone through our assessment phase. [7:02] This is a progressive progressive assessment that upgrades or updates as she progresses through our [7:07] AI readiness platform. In addition to that, she's also gone through our upskilling phase [7:14] where we can see that Maya has progressed in her learning journey and she's already completed her [7:20] cohort and now she's ready to unlock her a her capability journey. And that capability journey [7:27] happens in our capability studio. Now this is where as JPS mentioned we get to the point of [7:35] repeatability, duplicatability, understanding the real world impact. And so how do we do that? [7:42] We do it through a couple different ways. One of those ways is through AI adaptive simulations. [7:49] These AI adaptive simulations are realworld scenarios that Maya's going to be encountered [7:55] with in everyday workflows. And the and the purpose here is to build durable skills. [8:01] Skills in judgment, skills in adaptation, skills that she can turn into capability. [8:10] And now after she goes through these simulations, as JPS was referencing, she has the ability to [8:16] submit these with her reflections and evidence. And when she submits this, [8:24] uh, I'll wave here for a moment. I apologize. Uh when she puts in her reflections and her fe and [8:32] her feedback and how she's progressed through the simulation, it gets submitted to uh our skill up [8:39] mentors. Now, she'll get an instant realization of her assessment score from an AI and that AI will [8:49] uh then be sent for human review. Now, on to the next step, the projects as JPS was referring to. [8:56] So Maya's already completed a couple projects in her journey, but let's hop into one that she is [9:01] still working on. So as Maya goes through her project, she knows exactly what the purpose of [9:08] the project is. She knows her task in that project and she knows the success metrics [9:15] and along with that she has an AI project companion. Now this companion doesn't do [9:20] the work for Maya. What it does is it encourages her. It shows her the gaps of her reasoning and [9:27] thinking and it helps her build and iterate. And that build and iterate process is where the aha [9:32] moment begins to grow. Now, as Maya puts all of this down and she understands the exact impact, [9:40] she then provides the evidence for it. And this evidence is going to support her claims. [9:47] It's going to give the final results. And when she's done with that, she puts in her reflections, [9:55] uh what she how she felt about it, what she could improve, and where she thought something could [10:02] be changed. After that, she submits it and just like everything else on the AI readiness platform, [10:08] it's AI powered and humanled. And so that re that submission goes off for human review. [10:15] And this is where the aha moment really comes in. This is where the impact gets measured. And now [10:21] Maya can see that she's saved hundreds of hours peranom. She's had real world business impact [10:30] and she's built capability. And with that, I'll hand it over to Raton. Thank you, Ben, [10:37] for that wonderful demo. Uh there are two things I would like to highlight to conclude this demo. [10:44] The first as our learner Maya went through her process of AI capability building, [10:50] she went through the four-step process at scale up of assess, upscale, activate, and transform. [10:57] When she started building her AI capability, she didn't knew what she had in terms of cap AI [11:02] capability. So the assess helped her understand where she stood in terms of the capability her [11:07] manager, her company needed to be at. From there on she went through the upscale journey with us [11:14] where she had a custom cured a scaffolded pedagogy design real time into play. And from there on [11:21] once she had the knowledge and skills she went on to practice activate practice actual [11:27] hands-on projects to be able to transform the workflows in her in her business her actual work. [11:33] Now throughout this journey, Maya was Maya was helped by the Skillup AI models, [11:38] AI agents and AI systems. However, the secret sauce to Maya's success were humans at Skillup. [11:47] It were skill they were the skillup instructors, mentors and coaches who were working right [11:53] alongside with Maya to provide her the right guidance, the right support and most importantly [11:59] the right kind of AI orchestration she needed so that she can build the AI capability [12:05] in the most effective and efficient manner. This ties directly into the core philosophy [12:11] at scaleup when we are building AI capability in the world which is human-led AI powered. [12:19] The second important thing I want to uh highlight from the demo is that most AI trainings start and [12:26] end with hard AI skills alone. How to use the AI tools, how to write prompts, how to build agentic [12:35] AI workflows and so on. Now these skills matter but they are not enough. They are not holistic. [12:44] People also need durable skills to make sound decisions and work effectively with others. [12:50] And along with that a mindset to experiment, adapt and use AI responsibly. [12:57] So when Maya was building her AI capability with us at scaleup, she was constantly being trained, [13:03] pushed and improved on her durable skills as well. Her judgment improved with every project cycle she [13:11] did at Activate platform. She was forced to think critically with every iteration of AI that she ran [13:20] and she improved her verification skills to ensure that she followed the company defined AI safety [13:26] and standards. I submit to you that if not more durable skills and an adaptive mindset is equally [13:34] as important as the AI hard skills. Now we've been doing AI upscaling for over 10 years now. [13:41] We partner with enterprises, government and academy across the world to drive large, [13:45] small, medium, end-to-end upscale initiatives. But this year has been extra special for us. [13:53] Ever since we launched uh our AI readiness program, we've become super busy. [13:59] Right now, we have about 25 plus AI readiness programs running across the [14:03] globe with enterprises, government, and to make their workforce AI capable. And by the way, [14:09] that number just kind of moved to 32 is my what my team is telling me in the past two days. [14:14] But what's really heartening for us at scaleup is that not only these people are delivering [14:21] uh impact through ROI and productivity gains in their organizations, but they are [14:26] now doing it very confidently, comfortably with through a through true AI capability building. [14:35] And with that, I'd like to hand it back to JPS for closing comments. [14:39] Well, in the beginning I asked you a question. How many of you have completed an AI course? I want to [14:45] end with a [clears throat] different question. How many people in your organization can actually [14:52] demonstrate AI capability? And this is a really critical question because over the next decade, [15:01] this is just not a learning metric. This will become your competitive advantage. [clears throat] [15:09] Finally, at Scaleup, we believe human potential is still the world's greatest resource [15:16] and AI simply gives us a way to amplify it. And lasting value is not created by teaching more. [15:25] Instead, it is created by building capability that compounds and it transforms [15:33] through business impact and creating business value. Thank you very much. And [applause]