AI: A “Tool for Reducing Staff” or a “Device for Increasing Waste”?

Monday.com has laid off hundreds of employees and shifted its focus entirely to AI. In the same week, it was revealed that over 50% of daily uploads on Deezer are AI-generated songs, totaling more than 90,000 tracks in June alone.

When these two pieces of news are placed side by side, a pattern emerges.

Companies that have “structurally reduced labor costs” with AI versus companies that have “mass-produced worthless content” with AI.

Despite both utilizing AI, the outcomes are completely opposite. What determines this divergence? And where should small and medium-sized enterprises (SMEs) position themselves?

—

Monday.com Did Not Cut “People” but Rather “Structure”

Monday.com is a SaaS company specializing in project management tools. In July 2025, it announced layoffs affecting hundreds of employees and made it clear that it would reorganize around AI.

What is noteworthy here is that this is not a simple case of “cutting people because AI was introduced.”

What the company did was redesign its business processes with AI as the premise. They broke down workflows that previously required five people into tasks that AI could handle, leaving only the parts that humans should perform. As a result, the number of people needed to produce the same output was structurally reduced.

This is closer to a “design change” than a “restructuring.”

In SaaS companies, labor costs often account for 40-60% of revenue. If a company with an annual labor cost of 5 billion yen can transfer 20% of its operations to AI, it can save 1 billion yen annually. This saved capital can be redirected to product development or AI investment. While competitors are still relying on human labor, a significant cost structure advantage can be established.

Whether Monday.com’s decision is correct will be proven by its future performance, but at the very least, the fact that they clearly defined “where to use AI” and changed their organizational structure accordingly is on a different level from mere cost-cutting.

—

What’s Happening at Deezer — A World Where AI Mass-Produces “Cost-Free Waste”

On the other hand, the situation at Deezer is fundamentally different.

The music streaming platform is being inundated with AI-generated songs. Over 50% of daily uploads are AI-generated, with more than 90,000 tracks in June alone.

Why is this happening? The reason is simple: the cost of creating songs has approached nearly zero.

Traditionally, producing a single song required tens of thousands to hundreds of thousands of yen for composition, arrangement, recording, and mixing. With AI tools, however, a song can be created in just a few minutes by simply entering a prompt. The cost is virtually zero.

What happens when costs drop to zero? The “criteria for creation” disappears.

When humans create, the significant cost of tens of thousands of yen prompts serious consideration of whether a song is worth releasing. With AI, the cost is zero, leading to a mindset of “let’s just create and upload 100 songs for now.” If there’s a chance of earning a tiny revenue from streaming plays, it becomes rational to produce in bulk.

As a result, the platform becomes flooded with low-quality AI-generated content. Listeners struggle to find the songs they want. Artists’ works get buried. The value of the platform itself is compromised.

This is the structure of “increasing waste with AI.” The issue is not with AI itself, but rather with the lack of quality filters in a system where costs have become zero.

—

Substack’s Move — The Era Where “Human-Written” Becomes Valuable

In response to this trend, the newsletter platform Substack has made an interesting move by introducing AI detection tools.

They have implemented a feature that scans text of over 100 words to identify portions written by AI.

This is not just a technical response; it is a proactive move anticipating a market shift where “what is written by humans” becomes a differentiating factor.

As the cost of content generation approaches zero, AI-generated content will flood the market. In such a scenario, what becomes rare is “content that is thought out and written based on human experience.” Substack has labeled this rarity.

This structure offers extremely important insights for SMEs.

—

The True Divergence Point for SMEs

Now, let’s get to the main point.

Both the stories of Monday.com and Deezer may seem like tales of large corporations or platforms. However, the structures demonstrated by these two cases apply directly to local SMEs as well.

The Divergence Point is Summed Up in “Where to Use AI”

Pattern A: Use AI in the “Back Office” → Costs Decrease

• Creating estimates, processing invoices, aggregating daily reports, initial sorting of inquiries
• Accounting tasks that took 20 hours a month can be reduced to 3 hours
• The labor cost of one part-time employee (150,000 yen per month) can be replaced by an AI tool costing 5,000 yen per month
• The time and cost saved can be redirected to sales and customer service

This is the Monday.com approach. Hand over tasks that don’t need to be done by humans to AI, allowing humans to focus on what only humans can do.

Pattern B: Use AI directly for “Output” → Waste Increases

• Publishing blog articles generated by AI as is
• Sending sales emails generated by AI without modification
• Submitting proposals created by AI without review

This is the Deezer approach. Since the cost is zero, a large volume can be produced. However, the recipients can immediately tell. “Ah, this was written by AI.”

The strength of local SMEs lies in their “visible relationships.” The CEO speaks directly to customers, and those who understand the field make proposals. If AI-generated templates are layered on top of that trust, it will erode it. Costs may decrease, but the value diminishes even more.

The Criteria for Judgment is Simple

“Is the use of AI visible to the customer or not?”

• If it’s not visible to the customer (back office, internal operations) → Use it aggressively
• If it’s visible to the customer (proposals, content, communication) → Always pass it through a human filter

That’s all there is to it.

—

How SMEs Can Win in the “Zero Cost Era”

AI is dramatically reducing costs in content production, code generation, and design. This applies equally to both large and small businesses.

So how can SMEs differentiate themselves?

The answer is paradoxical: compete with “what AI cannot produce” rather than “what AI has made cheaper.”

• Long-standing relationships with local customers
• The judgment of humans who know the field
• Trust in being specifically chosen by someone
• Industry-specific tacit knowledge and experience

These are things that AI cannot generate. Moreover, with the spread of AI, the value of these aspects will increase relatively.

Let large corporations handle what AI commoditizes. SMEs should lower back-office costs with AI while concentrating management resources on values that only humans can provide. This is the third path that is neither about “reducing people with AI” nor “increasing waste with AI.”

—

So, What Should We Do?

Here are three actions you can start from tomorrow.

1. First, hand over one “repetitive task” in the back office to AI
Summarizing meeting minutes, drafting emails, data entry. You can start with tools costing a few thousand yen per month. Measure the effects numerically: “Hours saved per month” and “Yen saved per month.”

2. Ensure that all outputs reaching customers have human oversight and input
It’s fine to have AI create drafts. But the final check and finishing touches should be done by a human. The “AI-like quality” will be obvious to customers. The moment they realize it, trust will erode.

3. Articulate the “value that only we possess”
What are your company’s strengths that AI cannot replace? If this is not clarified, you won’t be able to determine which areas to invest in AI and which to invest in humans.

—

Monday.com changed its organizational structure. Deezer was overwhelmed by a flood of waste. Substack labeled “being human.”

While the responses of these three companies are varied, the questions they pose are the same.

“What costs will your company reduce with AI, and what value will it increase?”

Companies that cannot answer this question are at risk, regardless of which side they fall on.