This video features Shin Gunho, Head of B2B Business at DAY1COMPANY, presenting on the theme that AX (AI Transformation) is as difficult as founding a startup — but applying a startup framework can raise your odds of success. He shares lessons learned from missed AX opportunities and three playbooks for successful AX. First, take a thin but end-to-end approach to achieve fast PMF (Product-Market Fit). Second, pursue fundamental innovation through AI-native redesign. Third, become a Full-Stack AX Builder to gain speed. He argues that succeeding at AI transformation ultimately requires capabilities similar to those of an entrepreneur, and emphasizes that rapid execution and visible results are essential to winning ongoing investment from the CEO.
1. Shin Gunho's AX Experience and the Core Question
As Head of B2B Business at DAY1COMPANY, Shin Gunho has been deeply focused on building AX (AI Transformation) with revenue generation as his north star. He introduces himself as a consultant with a broad but shallow skill set who built an engineer's mindset through Python-based modeling, machine learning, and deep learning in the financial sector. Today he provides AX training to enterprise clients, gaining insight into the diverse AX needs of large organizations, and is preparing to launch a new business that combines consulting and build services.
Internally, he built and deployed a variety of AX projects in 2025 — GPTs, dashboards, agents — to improve productivity, and formally established an AI Transformation team in November 2025. Despite limited resources, the team selected the four sharpest people from the training team, had them learn n8n, and is now completely rebuilding the proposal workflow. He reports the results have exceeded expectations.
Through his presentation, Shin raises the central questions: "How do we make AI transformation succeed?" and "How can we go beyond individual productivity to organization-wide productivity gains?" — inviting everyone in the room to recognize they share the same challenge. 🗣️ He emphasizes that on the technology adoption curve, most internal business members at large companies fall into the late majority or laggards rather than innovators or early adopters, making organization-wide AX success an extremely difficult undertaking — even leadership typically gets only two or three chances to attempt it.
"When I'm in AX meetings and training meetings with large enterprises, and I look at the internal business members — the people in this room are about here. You're wildly extreme outliers. Where is my team lead? If the team lead is somewhere around here from a business perspective, I'm truly grateful. And if the marketing team, training team, and sales team that need to do AX are somewhere around here, I'm truly grateful."
"Ultimately, to get all the way to organizational productivity, you have to force these people. I have to do it. I have to get all of this done with these people — and the usual response is: 'Why should I be involved with you?'"
2. Lessons from AX Failures and Successes: The Importance of 10x Value
Shin Gunho shares his experiences with AX successes and failures, and the important lessons he drew from them.
2.1. Failed Case: Building a Proposal Search System 📉
Outside of the training business, he launched his first AX project with the idea of building a "proposal search system." He analyzed the workflow in which training consultants listen to client needs and recommend training programs, built four agents, and deployed the system — only for it to fail. He says it now sits unused and abandoned.
Analyzing the failure, he found the system reduced proposal selection time from one to two hours down to fifteen minutes — roughly a 4x efficiency gain — but this was not enough, because it failed to deliver sufficient value to make users switch.
"I built this proposal search system and deployed it. And it failed. Nobody uses it now. It's just sitting there, abandoned."
2.2. Successful Case: AX for the Skills Consulting Pipeline 🚀
By contrast, his second attempt — AX for the "skills consulting" service — was a major success. Skills consulting is a complex process: defining the software skills needed across 80 teams and roughly 100 job functions in the R&D organization of a battery solutions company, then building a training system around them. It was a massive project where even four people working overnight for two weeks wouldn't be enough.
Shin approached this project using "vibe coding" — skipping formal workflow analysis and jumping straight to writing logic and building agents all the way through to generating interview sheets. He worked with GPT to structure the question logic a consultant would need and developed it himself, completing the full AS (Automation System) pipeline for this complex consulting workflow through five iterations.
"I built the entire pipeline for this difficult consulting work — the full AS pipeline — and ran five iterations on it myself."
"I'd write out the logic, check the output all the way through, think 'the logic is too heavy,' go back and start from scratch, all the way to the end again. What was truly remarkable was that in traditional consulting you can't run iterations like this — unlike software development, people have to grind for a very long time and only see results eight weeks later. But now, working with AI, I ran this five times and then deployed it to the team."
He invested five hours every weekend to build this workflow, which connected five GPTs passing JSON data back and forth to ultimately generate Excel-formatted interview sheets. It was a far less familiar method than the previous proposal search system, yet every team member who needed to do this work switched to the new workflow and actually used it.
"People who don't know anything — people who don't even know what JSON is — when you just send them a link to a web page and say 'use this,' they absolutely will not. But everyone used this so much that they all switched to the new workflow."
2.3. The Decisive Difference in AX Success: 10x Value
From these two cases, Shin emphasizes the key insight he gained: the decisive factor in AX success is the magnitude of the efficiency gain. A 4x improvement is not enough for users to overcome the switching cost of abandoning their existing approach for a new AX system. The skills consulting AX, which dramatically cut a task that previously took 120 hours and made it possible to "escape somehow," succeeded precisely because it delivered 10x or more value.
"The decisive difference came down to exactly one thing. The first lesson I learned: the size of the efficiency gain determines whether AX succeeds."
"The efficiency level I've always had in mind is 10x or 20x — at that level, the switch is guaranteed. Four or five times is nowhere near enough. That's what I came to understand."
3. AX Is No Different from Founding a Startup: Applying the Startup Framework
Shin Gunho says he arrived at the conclusion that sufficiently advanced AX is ultimately no different from founding a startup. Just as a startup must deliver 10x value to customers through a new product in order to succeed, AX must also generate 10x value to get organizational members (internal customers) to overcome their switching costs and adopt a new system.
"Ultimately, to overcome that switching cost you need to reach that level. Doesn't that sound a lot like a startup?" "Whether it's a startup or AX, from the perspective of delivering a product that creates 10x value for the customer, you can apply the startup framework in almost exactly the same way."
He maps the startup framework onto AX as follows:
- Market 📈: Just as a startup must choose a large market, in AX the size of time spent per organization or workflow serves as the market size. With limited resources and leadership authority to protect, you must choose a "market" with high switching volume.
- MVP 🛠️: Like a startup's MVP, AX goes through a PoC (proof of concept) for validation before scaling up.
- PMF 🎯: Just as a startup finds PMF when customers pay money, in AX you find PMF when internal customers (organizational members) become so dependent on the new system they can't go back to the old way. This means designing the product so that even marketing team members who know nothing about AI naturally think in terms of the new user flow and follow along.
- Investors 💰: If a startup's investors are VCs, AX's investors are senior decision-makers. You must prove ROI on resources invested to earn more investment.
- Runway ⏳: Just as a startup goes bankrupt when it runs out of cash, for an AX team the runway is the trust of senior decision-makers. When trust runs dry, authority is revoked — organizational bankruptcy.
- Difficulty 🤯: Everyone knows startups are hard, but Shin says AX carries the same level of difficulty. He even quips, "At least if you just founded a startup you could make some money," underscoring how hard AX really is.
"PoC is the point where the real work begins. So of course, a successful PoC absolutely does not guarantee AX success."
4. The Capabilities Needed for AX Success: The Full-Stack AX Builder
If an AX builder is like an entrepreneur, what combination of capabilities does AX success require? Shin argues that beyond existing Echo and Delta capabilities, you need the combined capabilities of a PO (Product Owner) and PM (Project Manager) — what he calls "Full-Stack AX" capability.
- Consultant capabilities 🧠: The ability to quickly decompose and structure problems, identify bottlenecks, and iterate on logic. Especially important is ultra-fast execution — rapidly checking results and, on failure, quickly returning to try again.
- Product Owner (PO) capabilities 🙋♀️: Understanding human laziness and boredom; a sense for products people will actually use; a concept of iteration; and an obsession with the problem.
- Project Manager (PM) capabilities ⏱️: Persuading uncooperative stakeholders throughout the AX process, managing plans within constrained timelines, and expanding from PoC success to major projects — requiring schedule management and resilience.
He emphasizes that all of these capabilities are no different from what an entrepreneur must possess.
"Ultimately, the mindset you need for AX is this: entrepreneurs think 'nothing is impossible in this world — I just haven't found the way yet.' Now it's 'nothing is impossible for AI — I just gave the wrong instructions.' Holding onto that mindset with tenacity is the capability set I believe you need to succeed."
5. Three Playbooks for AX Success
Shin Gunho presents three playbooks for achieving AX success.
5.1. Playbook 1: Achieve Fast PMF with a Thin but End-to-End Approach 💡
The first playbook is a "thin but end-to-end" approach. In business reality, not all organizational members follow the same workflow — everyone's workflow is different. Targeting only the bottleneck portion of a specific workflow — a "thick" partial approach — addresses only a small market, making organization-wide AX success difficult.
By contrast, the "thin but end-to-end" approach doesn't trust the workflow itself; it focuses on the fact that every workflow ultimately shares the same inputs and outputs. If everyone needs to use the same input in their own way to produce the same final output (e.g., a proposal), AX-ing that entire process end-to-end is more likely to lead to organization-wide AX adoption.
Going thin also dramatically speeds up iteration cycles, making it possible to discover and fix real problems faster. He warns that trying to build too thick and too perfectly risks burning out before the MVP even launches.
"Going thin but end-to-end is more likely to achieve PMF quickly. That's the first playbook."
"A month passes and someone asks, 'Did you get anything done?' — 'Did you make any money?' — that's what they're asking. In that sense, going thin is what keeps the iteration loop spinning fast, and that's where you're more likely to discover the real problem."
5.2. Playbook 2: Create 10x Value through AI-Native Redesign 🚀
The second playbook is "building AI-native from the start to achieve 10x value." He explains that AX requires applying First Principles Thinking — decomposing complex problems down to their most fundamental level, à la Elon Musk — rather than simply following existing workflows because someone said the work is heavy or hard.
Rather than following existing workflows as-is because people complain that they're overloaded, you should focus on the essence: every intellectual activity is a process of converting low-value information into high-value information. Decompose the stages in which information value increases, and ruthlessly eliminate the steps that can be delegated to agents (AI).
He emphasizes that questioning tasks humans have always taken for granted, erasing everything, and redesigning AI-natively makes it far easier to achieve 10x efficiency. This means not merely automating parts of an existing workflow, but completely rethinking processes from an AI perspective.
"Building AI-native from scratch to get to 10x is a lot easier."
"Ultimately, everything we do is an act of converting low-value information into high-value information. That's all any activity is."
"Just erase everything and go AI-native — getting to 10x is much easier that way. That's the second playbook."
5.3. Playbook 3: Gain Speed by Becoming a Full-Stack AX Builder ⚡
The third playbook is "you need to become a Full-Stack AX Builder." He emphasizes that ultimately the most important factor in AX success is speed. Before the trust and patience of senior decision-makers runs out, you must quickly produce visible results and change — only then will you continue to receive investment.
He shares that he himself is working to combine Echo (consultant) and Delta (developer) capabilities, and closes his talk by saying that the frontend-focused engineer developers in the room must also continuously expand their capabilities and skill sets — from Delta toward Echo, and further toward full-stack — because only then can they gain the speed needed to make AX succeed.
"Ultimately, before everything solidifies — before the scales of trust and patience tip to zero — what it comes down to is speed."
"From an AX perspective, you need to become full-stack to succeed."
Closing
Shin Gunho makes it unmistakably clear that AX is a challenge on par with founding a startup — far beyond simple technology adoption. His message resonates deeply for anyone grappling with AI transformation in their organization: you must deliver revolutionary 10x value, apply an agile startup-like framework, and prove results fast as a full-stack builder who combines the capabilities of consultant, PO, and PM. What matters in the end is not AI technology itself, but the innovator's mindset and execution power that fundamentally changes the way people work — and with that, he concluded his talk. 👏