The unfortunate reality that the more a company studies AI, the less they use it.
Hello, I am
Yoshiki Fukiage, CEO of Lean Stack Inc..
[Author Profile]
Yoshiki Fukiage
Representative Director, Lean Stack Inc.
April 2021: Joined the Ministry of Economy, Trade and Industry (METI)
July 2023: Resigned from METI
September 2023: Founded Lean Stack Inc.
Company Website: https://www.leanstack-buzz.com/
YouTube Channel: DX Dekiru-kun
https://youtube.com/channel/UC-f1ZI9MiME-GC0l1gwNf4w?si=WeyqGjBSvVXwSaDB
Is your company still "studying" AI today?
You attended a seminar.
You bookmarked an article.
You compiled a comparison table in a spreadsheet.
You shared it on the company Slack saying, "This is helpful!"
Congratulations.
So, did your business performance improve by even one yen?
It didn't, did it?
I know.
I have worked with dozens of companies in the field of DX support, and it is overwhelmingly common that the companies that proudly claim, "We are studying AI so much!" are the ones that haven't earned a single yen from AI.
That is not sarcasm; it is a fact.
Today, I will talk about this structure without mercy.
Here we go.
The problem where nothing changes even though there are "people knowledgeable about AI" in the company
Recently, I often encounter companies like this.
The president is constantly attending AI seminars.
Young staff in the IT department are using ChatGPT extensively.
Corporate planning is compiling an "AI Utilization Report" every week.
And so, I ask.
"In reality, what has changed in your operations?"
The response you get is usually like this.
“...We are currently in the middle of considering it.”
“Under consideration.”
This phrase might be the most used Japanese expression in the AI industry.
In short,even though there are people who are knowledgeable, it ends with just being knowledgeable.
Why does this happen?
The answer is simple.
“Knowing” and “using” are completely different muscles.
Being well-versed in the latest AI trends has absolutely nothing to do with shortening the time it takes to process an invoice in front of you by 10 minutes.
Just because someone can compare the performance of GPT and Claude doesn't mean they can improve their company's workflow.
In fact, the more you know,
“Maybe there's a better tool”
“Maybe it's still too early”
you become cautious and end up stuck forever.
AI knowledge is just trivia if you don't use it.
It might make for good conversation at a drinking party, though.
The productivity killer known as the “AI Study Group”
The thing I think is most dangerous lately.
In-house AI study groups.
I might get in trouble for saying this, but I will.
In most companies, that has becomea massive waste of time.
Isn't the typical flow something like this?
Someone makes slides like “AI Basics”
Ten people gather in a conference room.
Explaining "What is ChatGPT?" and "What is prompt engineering?"
During Q&A, someone asks, "Can we use this for our work?"
Replying, "It depends on how you use it."
Everyone leaves feeling somewhat smarter.
The next day, everyone does their usual work as usual.
Yes, nothing has changed.
This is just cosplay in the name of training.
A pose of "We are working on AI." Creating a track record to write in the management plan.
Spending 10 people x 2 hours = 20 hours of employee time, with zero results produced.
If you had 20 hours, you could have run three experiments to automate monthly accounting processes with AI.
I'm not saying you should stop the study sessions. But I do want to ask: aren't you just intoxicated by the fact that you're holding study sessions?
People who chase information are not "running," they are "spinning."
The trouble with gathering AI information is that
it gives you the feeling that you're doing something.
If you go to a seminar, you feel like you've "learned." If you read an article, you feel like you've "caught up." If you start a tool trial, you feel like you're "moving toward implementation."
But I want you to think about it calmly.
How is that any different from a hamster wheel?
You're running hard. You're sweating. You're tired.
But you haven't moved forward even a millimeter.
The more information you gather, the more options you have. The more options you have, the less you can decide. Because you can't decide, you gather more information.
The infinite loop is complete.
And what's more, it's not the lazy people who get stuck in this loop, but the most diligent ones.
"I want to research it thoroughly before I act."
"I don't want to make a bad investment."
"I want to make the optimal choice."
This diligence is acting as the biggest brake.
Cruel, isn't it?
So, what do we do about it?
Since I've spent so much time criticizing, it's time to offer a solution.
What you need to do is incredibly simple.
Tomorrow, go around the office and ask people what their most 'tedious task' is.
Don't mention AI at all.
Don't name any tools.
Absolutely do not use the word 'DX'.
Just ask.
"What's the most annoying task you have right now?"
That's it.
It might be copy-pasting invoices.
It might be transcribing daily reports.
It might be the report you have to create in the same format every month.
Once that 'annoying' task is identified, throw AI at just that one thing.
ChatGPT, Claude, anything is fine.
Something you can try for free is more than enough.
The starting line for AI utilization isn't catching up on the latest information.
It's being able to honestly say, 'What is tedious about our work on the ground?'
Just that.
Seminars, study groups, and comparison charts—you don't need any of those right now.
Start by finding the tedious tasks within your company.
Really, start here.
Finally, one thing I want to say in all seriousness.
Some of you reading this might think, 'Hey, I don't want to be criticized by you.'
But look, I'm not making fun of studying AI.
Since I make my living in this field, I follow AI information every single day.
And, to be honest,
even I cannot keep up with everything.
That is exactly why I can say this with certainty.
'Trying to know everything before acting' is synonymous with never acting at all.
Ironically, the more you know, the less you can move.
A company that acted when its knowledge was at 1 will have 100 in experience by the time its knowledge reaches 10.
A company that waited until its knowledge reached 10 will find that the market has already changed by the time it gets there.
A company that waited until its knowledge reached 10 will find that the market has already changed by the time it gets there.
Only companies that take action can find the next correct answer.
So, enough with the studying.
Tomorrow, go to the front lines and ask.
'What's a pain in the neck?'
Within that answer lies the correct way for your company to utilize AI.
I am convinced of this.
That is all for today.
Thank you for reading until the end.
Look forward to the next article.
[Author Profile]
Yoshiki Fukiage
CEO, Lean Stack Inc.
April 2021: Joined the Ministry of Economy, Trade and Industry. July 2023: Resigned from the Ministry of Economy, Trade and Industry. September 2023: Established Lean Stack Inc.
Our website: https://www.leanstack-buzz.com/
YouTube channel: DX Dekiru-kun
https://youtube.com/channel/UC-f1ZI9MiME-GC0l1gwNf4w?si=WeyqGjBSvVXwSaDB
Please feel free to contact us here
↓
https://lin.ee/I73KUiN