PCOSSAM PCOSSAM
LECTURE PREVIEW

Watch just 1 minute of the lecture
and you can decide whether to book it

8 clips filmed live at a corporate special lecture, ‘AI-Friendly Work Innovation.’ The slide structure, case selection, and delivery style are exactly as they happened. Below each clip, you'll find the full content written out in text.

8 clips · ~6 minutes total Kangwon National Univ. UICF · K-Hi-Tech Platform special lecture 200+ lectures/year · 5,000+ trainees/year
How to use this page — If you're a training coordinator, watch the clip on a topic you're interested in, then use the ‘Ask about this topic’ button below it to request a quote right away. Every lecture can be restructured to fit your organization as a 60-minute, 90-minute, 4-hour, or full-day course.
EPISODE 02 · Data Usage

Why Does China Use Data by the Second?

While Korean companies run quarterly consumer surveys, they collect reactions by the second.

A Chinese C-beauty brand analyzes consumer comments posted during livestream-commerce broadcasts in real time. Because the comments are already text, they can be analyzed immediately, with no extra processing. The resulting insights are applied to the product's formulation and packaging design right away.

Cosmetics have moved from being a ‘manufactured good you make once and sell’ to something closer to ‘software that keeps updating based on feedback.’ This is what AX (AI Transformation, an AI-driven business shift) actually looks like. The speed gap becomes the product's competitive gap.

  • Only data that accumulates as text can be analyzed in real time
  • A decision-making structure where analysis results feed straight into formulation and packaging
  • The shift from simple manufacturing to a feedback-driven, software-update model (AX)
Ask about this topic Custom AI Transformation Partnership →
EPISODE 03 · Prompting in Practice

When You Summarize an Article with AI, Ask for Examples, Not Summaries

“Summarize this” is one of the worst prompts you can write.

The vaguer the request, the more generic and useless ChatGPT's answer becomes. Instead, attach the article PDF and specify the shape of the output — for example, "what is this article actually arguing, and what are its 3 key keywords?"

Asking specifically for examples, rather than a summary, leaves you with concrete, quotable scenes instead of an abstract paragraph. What actually gets used in reports and presentations isn't the summary — it's the example.

  • Specify the shape of the output first, instead of a vague "summarize"
  • Pin down scope with a number, like "3 key keywords"
  • Asking for examples leaves you with sentences you can use directly in a presentation
Ask about this topic Generative AI Training →
EPISODE 04 · AI Strategy

Why It Helps to Build a GPT Clone in Advance

The key is building it ahead of time, not the moment you need it.

Use the ChatGPT Projects feature to build an ‘Elon Musk’ clone in advance. With his statements and way of thinking already trained into it, asking “how would Elon Musk see this problem?” returns an answer from that vantage point.

Place a Jensen Huang clone next to it and ask the same question, and the difference between the two viewpoints becomes obvious. When reviewing a proposal or pitch deck, this approach functions as a standing advisory board on permanent call.

  • Build a project per person and train it continuously on their statements and thinking
  • Use "if you were the expert" perspective-shift questions to stress-test a proposal
  • Pair different persona clones together to surface the gap between viewpoints
Ask about this topic AI Presentation & Pitch Deck Consulting →
EPISODE 05 · Persuasion & Data

Why Did Amazon Ship Unpurchased Items?

Two out of every three people just kept the item. This wasn't a system built from data alone.

Amazon would ship an item the customer never ordered, with a note saying “if you don't want it, return it for free.” The starting point of this anticipatory shipping system wasn't an algorithm — it was people.

Amazon first brought in experts in psychology, behavioral economics, and consumer behavior to define ‘under what conditions does someone give up on returning an item,’ and only then trained purchase data on top of that foundation. It's expert input plus data training — as a combination. Organizations where AI adoption fails usually only do the second half.

  • Pre-registered payment methods → a frictionless, instant-purchase structure
  • A two-layer structure: experts set the hypothesis, data refines it
  • Persuasion starts with understanding human behavior first, not with technology
Ask about this topic Presentation & Pitch Deck Coaching →
EPISODE 06 · Report Writing

Why You Need to Write Reports AI Can Read

The next reader of the report you're writing now isn't your team lead — it's an AI, six months from now.

Create a project in ChatGPT for something like ‘Work Innovation’ and pile the related PDFs into one place — from that moment, your documents become a searchable asset. Instead of digging through files every time you need something, you can pull it out with a single question.

On the other hand, AI can't read a scanned image of a report, or a table that's actually just a picture. Just following three rules — a consistent heading structure, text-based tables, and a file-naming convention — completely changes how reusable your documents are.

  • Group documents into topic-based projects to turn them into a searchable asset
  • Scanned images and picture-based tables can't be read by AI → write everything as text
  • Heading structure, table format, and file-naming rules determine how reusable your documents are
Ask about this topic AI Report & Proposal Writing Training →
EPISODE 07 · AI Trends

AI Learns from Data, Robots Learn from Action

The next round won't be fought over language — it'll be fought over behavioral data.

In Shanghai, China, there's a humanoid robot school, effectively a ‘robot kindergarten.’ It repeats 15 tasks — stacking boxes, hammering, climbing stairs, jumping rope, typing — to collect behavioral data.

Text data is already piled up across the internet, but behavioral data — things like ‘at what angle does a hand grip an object’ — simply doesn't exist yet in the world. So it's being generated and accumulated on purpose. It's time to ask what your own organization's ‘behavioral data’ is, and where it's being captured right now.

  • Language data already exists; behavioral data has to be generated directly
  • A structure that accumulates individual behavioral data through 15 repeated tasks
  • “In what form is our organization's know-how being captured right now?”
Ask about this topic Vibe Coding Hands-On Course →
EPISODE 08 · AI & Search

Why Did Google Digitize Every Library Book in the World?

While people gathered to search, the data was quietly accumulating in the opposite direction.

Over 15 years, Google scanned 30 million volumes from libraries around the world. At the time, it looked like a project purely about search convenience.

In the end, it turned out to be the work of converting human expert knowledge into a machine-readable form, and it became the training foundation for today's large language models. The same question needs to be asked of every corporate website right now: does our company's information exist in a form AI can read?

  • 15 years, 30 million books scanned → converted into machine-readable knowledge
  • A data-accumulation strategy hidden behind the surface goal of search convenience
  • Corporate websites also need to be redesigned around an AI-readable structure (AEO/GEO)
Ask about this topic AEO/GEO Structure Diagnosis & Improvement Consulting →
EPISODE 09 · Content Trends

How Did People Spend Their Time Over the Last 100 Years?

Every turning point has always been technology. Right now, the language barrier is the one coming down.

There's data tracking who people spent their day with, from 1930 through 2024 — family, school, friends, neighbors, church… and then, finally, ‘online’ shows up. The spread of Windows, digital cameras in the 2000s, and the iPhone in 2007 each created their own turning point.

In the text era, content could only be shared among people who understood the same language, but photos and video worked even without a shared language. Right now, generative AI and real-time translation are tearing down the next barrier.

  • Every new technology reshaped how people spent their time
  • The shift from text to photo and video steadily lowered the language barrier
  • The size of the next market that generative AI and real-time translation will create
Ask about this topic AI Marketing Website Building Course →

We'll rebuild this lecture to fit your organization

Even the same topic is designed completely differently depending on audience and time available. From a 60-minute special lecture to a 2-day intensive course — tell us your training goal, and we'll reply with a curriculum and quote within 1–2 business days.

Ask About Training See Services & Pricing