inflearn logo

How to determine whether an AI chatbot will be adopted or discarded

Have you ever seen a 200 million won chatbot being discarded? Between 70% and 85% of AI projects fail to achieve their expected results. In manufacturing, this number hits even harder. A chatbot that looked perfect in a vendor demo starts confidently giving wrong answers once actual field documents are fed into it. Skilled workers with 30 years of experience turn away, saying, "That thing is useless." Systems built with hundreds of millions of won are quietly abandoned. For over 16 years, I have personally planned and executed MES, Smart Factory, and Manufacturing DX/AX projects at Samsung Electronics, Samsung Display, Hyundai Mobis, and LS ELECTRIC. Now, I am designing and building AI chatbots specifically for manufacturing sites. Through this process, I have constantly pondered, "Why do so many manufacturing chatbots fail?" and I have reached the conclusion that the answer lies not in the technology, but in the planning. This course does not teach code. This is not a course for developers who build chatbots. It is for those who must plan the introduction of a chatbot, evaluate vendors, verify performance, and ultimately decide whether to adopt the system. It is designed for planners, managers, and team leaders in the manufacturing field. Without coding, in just 50 minutes, you will be able to: 1. Understand core AI chatbot technologies (RAG, Knowledge Graph) through simple analogies. 2. Immediately use 5 practical questions to distinguish a vendor's true capabilities. 3. Design a 30-day Proof of Concept (POC) using a single-page checklist. 4. Establish strategies to integrate chatbots without resistance from field workers. Why is this course necessary? - While the AI adoption rate in the Korean manufacturing industry remains at 24%, 81% of companies plan to expand their AI investment. This means that while adoption is low, the will to implement it is high. The problem is a lack of planners who know how to execute. - There are courses on Inflearn regarding RAG and chatbots, but they are all aimed at developers. This is the first AI chatbot planning course specifically for non-developers in the manufacturing industry. 50 minutes can prevent a 200 million won mistake.

1 learners are taking this course

Level Beginner

Course period Unlimited

Chatbot
Chatbot
RAG
RAG
knowledge-graph
knowledge-graph
Chatbot
Chatbot
RAG
RAG
knowledge-graph
knowledge-graph

What you will gain after the course

  • Judgment that won't be fooled by vendor demos — even without technical knowledge, you will be able to distinguish a vendor's true capabilities and pinpoint flaws in their proposal with just five questions.

  • 30-Day POC Design Capability — From selecting use cases to defining success criteria, field testing, and making Go/No-Go decisions, you will be able to design a pilot implementation yourself with a single checklist.

  • Understanding core AI chatbot technologies for non-developers — You will be able to explain the difference between RAG and Knowledge Graph with a single analogy and judge for yourself which method is right for your factory.

  • On-site implementation strategy — You will be able to immediately apply practical positioning methods to turn the resistance of skilled workers into support, along with introduction strategies that start small and scale up.

🏭 How to Plan a Manufacturing Chatbot to Avoid a 200 Million Won Mistake

Have you ever had an experience like this?

"Let's implement an AI chatbot too." This directive came down from above. So you called in a vendor, and their demo looks quite impressive. Yet, a part of you feels uneasy.

"Will this actually work in our factory?" "Is it okay to sign the contract just by trusting the vendor's words?" "Who will take responsibility if it fails after implementation?"

This anxiety is well-founded. 70–85% of AI projects fail to achieve their expected results, and 42% of companies have abandoned their AI initiatives altogether. In manufacturing, it is a reality that chatbots costing hundreds of millions of won are being quietly discarded.


🔍 Why are there so many failures?

It's not because the technology is lacking. It's because the planning was flawed.

  • The hallucination problem, where AI confidently makes up things it doesn't know, was not verified in advance.

  • They did not consider the Korean-English mixed-use issue where "작업지시" and "Work Order" have the same meaning but cannot be searched.

  • We judged based only on the demo prepared by the vendor and did not test with our own documents

  • We did not manage the sense of threat felt by skilled workers with 30 years of experience.

  • We did not check who has data ownership after the contract ends.

All of these failures could have been prevented before implementation—if only you had known the right questions to ask.


📋 What this lecture covers

Lecture 1. Why 70% of Manufacturing Chatbots Fail (Free Access) Analyzes 5 structural patterns of manufacturing chatbot failures. It demonstrates through numbers and case studies the core message that these are planning issues, not technical ones.

Lesson 2. RAG vs. Knowledge Graph, Explained Through Metaphors Explains the two core technologies of AI chatbots using the metaphors of "library search" and "factory organizational chart." We provide criteria for determining why problems occur in mixed Korean-English documents at manufacturing sites and which method is suitable for your factory.

Lesson 3. 5 Questions to Avoid Getting Fooled by Vendors We provide 5 killer questions that allow you to distinguish a vendor's true capabilities even if you don't have technical knowledge. For each question, we specifically show the difference between a good answer and a bad one.

Lesson 4. How to Design a 30-Day POC Design a week-by-week 30-day pilot introduction to test if it actually works before full-scale implementation. Along with the core principle of "don't let the vendor choose the scenarios," it covers why triangular collaboration between the planner, internal system manager, and vendor is essential.

Lecture 5. Three Conditions for Chatbots to Survive in the Field Even if it passes technical verification, it is meaningless if it is not used in the field. This session covers how to make skilled workers your allies, strategies for starting small and proving results quickly, and a structure where AI suggests while humans make the final judgment.


👤 Instructor Introduction

I am a current manufacturing DX practitioner who has personally planned and executed MES, Smart Factory, and manufacturing DX/AX in manufacturing fields for over 16 years, including at Samsung Electronics, Samsung Display, Hyundai Mobis, and LS ELECTRIC. I studied the intersection of technology and management during my Master's program in Management of Technology at KAIST, and I am currently leading projects to directly design and build AI chatbots for manufacturing sites.

This lecture is delivered from the client's perspective rather than the vendor's, and is based on field experience rather than theory.


✅ Recommended for the following people

  • Those who have been instructed to "look into implementing a chatbot" but have no idea where to start

  • Those who have seen vendor demos but aren't sure if it actually works

  • Those who have introduced an AI chatbot, but find it is not being used in the field

  • Those who want to take the lead as a planner without being dragged around by technical jargon

❌ This course is not suitable for the following people

  • Those who want to build an AI chatbot by coding it themselves (There is no code in this lecture)

  • Those who want to learn development frameworks such as RAG/LangChain

  • Those who are planning chatbots for fields other than manufacturing


💡 Notes before taking the course

  • No prior knowledge is required. Knowledge of coding, AI, or programming is not needed at all.

  • There are a total of 5 lessons, lasting about 50 minutes. It is a length that allows you to listen to one lesson at a time during your lunch break.

  • The first lesson is available for free, so you can listen to it first before making a decision.

Recommended for
these people

Who is this course right for?

  • Manufacturing planners or managers who have been instructed to "look into implementing a chatbot" but don't know where to start — those who feel overwhelmed by the situation of having to evaluate vendor proposals without a technical background.

  • Those who have seen the vendor demo but aren't convinced if "this actually works"—those who feel frustrated because the screen looks plausible, but they have no criteria to judge whether it will work just as well with their own factory documents.

  • Those who have already introduced an AI chatbot but find it unused in the field — those whose systems are running but ignored by workers, and who are struggling because they cannot prove performance relative to investment.

  • Those who are in charge of Smart Factory/Manufacturing DX but feel like they are being led by the development team or vendors in AI-related decision-making — those who want to take the lead as a planner without being overwhelmed by technical terminology.

Need to know before starting?

  • Those who have experience in manufacturing field operations (factory, production, quality, facilities, etc.)

  • Those who have at least heard of manufacturing systems like MES or ERP.

  • Anyone who has used AI services like ChatGPT or Claude at least once

Hello
This is fleem826937

38

Learners

5

Reviews

4.4

Rating

3

Courses

With over 15 years of experience in production technology and equipment engineering within the manufacturing sector, I have specialized in solving on-site challenges through data and systems. Starting with PC-based equipment control, I have built my expertise in systemic improvement by understanding process and equipment structures and analyzing manufacturing data flows and operational frameworks.
Currently, I design and implement practical solutions in the field of Manufacturing AX (AI & Digital Transformation), connecting data, processes, systems, and automation.

www.linkedin.com/in/기호-이-3015a317b

More

Curriculum

All

5 lectures ∙ (41min)

Course Materials:

Lecture resources
Published: 
Last updated: 

Reviews

Not enough reviews.
Please write a valuable review that helps everyone!

fleem826937's other courses

Check out other courses by the instructor!

Similar courses

Explore other courses in the same field!

Limited time deal

$15,400.00

30%

$17.60