Tech talent/freelance market outlook in the AI era | Recruitment platform and agent structure

[Summary]

With the spread of AI, the tech talent market is not simply moving in a direction where people are no longer needed.

Rather, what is happening is a reshuffling of demand.

Work such as simple implementation, routine testing, document creation, and screen mass production can easily be compressed by AI. On the other hand, the value of human resources who can use AI to advance business design, data coordination, security, quality verification, and productization is increasing.

Documents from the Ministry of Economy, Trade and Industry have long referenced an estimate that there will be a shortage of up to 790,000 IT personnel in 2030. However, as of 2026, investors should be looking at more than just a "lack of staff." The problem lies in the nature of the shortage of human resources.

In the human resources market in 2026-2027, we expect to see a shift from a model where job advertisements are posted and people wait, to an infrastructure-based model where skill data, work history, learning history, and candidate contact points are analyzed using AI and linked to recruitment and development results.

Recruit HD (6098), Persol HD (2181), Visional (4194), and Wantedly (3991) each approach this theme from a different angle.

What is even more difficult to overlook is the market for IT freelancers, outsourced work, and side job personnel outside of full-time employment. It has become difficult for companies to hire full-time AI personnel, and they are increasingly procuring specialized human resources from outside when needed.

However, not all companies win equally. What is at stake in the AI ​​recruitment/freelance market is not the number of registrants, but the quality of data, points of contact with companies, shallowness of commercial channels, monetization, and the operational ability to bring results to recruitment and development.

AI will not eliminate the demand for human resources, but will replace it

When thinking about the tech talent market in the AI era, the first thing to differentiate is "workload" and "human resource value."

As code generation AI becomes more widespread, the amount of simple implementation will decrease. This is quite realistic.

• CRUD implementation *Standard test
• SQL generation
• Draft specifications
• API template creation
• Document maintenance

These tasks can be speeded up with AI assistance. From a company's perspective, there are times when the number of people needed to produce the same product is reduced.

But that doesn't mean the demand for engineers will disappear all at once.

There are actually not enough people to review the code created by AI, to break down business requirements, to connect with existing systems, to design security and privilege management, and to take responsibility for operations monitoring.

In other words, demand changes like this.

Key point
        ↓
AIKey point

This is a structural shift in the recruitment market.

The IT talent shortage in 2030 will be a matter of “quality” rather than “quantity”

A frequently quoted document from the Ministry of Economy, Trade and Industry is an estimate that there will be a shortage of up to 790,000 IT human resources in 2030.

This number itself is an estimate that has been used for a long time, but it is still often referenced as a premise for the recruitment market.

However, if we look at the situation in 2026, it is insufficient to simply say that there is a shortage of IT human resources.

What is missing is not so much programming experience.

• People who can incorporate AI utilization into their work
• People who can design data infrastructure
• Person who can verify errors in AI output
• People who understand security and governance
• People who can translate between field operations and systems
• People who can see operational KPIs after introducing AI

It is.

IPA's DX Trends 2025 also shows that the lack of human resources to promote DX is a major issue for Japanese companies, and that there is strong demand in the areas of AI, big data, and cybersecurity.

What is missing in the AI ​​era is not people who can use AI tools.

A person who uses AI in business and turns it into results.

Recruitment market shifts from job advertisements to AI matching infrastructure

In the traditional recruitment market, job advertisements, career change sites, human resources introduction, temporary staffing, and scouting services each played a role.

However, for tech and AI talent, it is becoming difficult to just post a job posting and wait.

There are three reasons.

1. The more talented people are less likely to enter the job market
2. Skills change rapidly and are difficult to evaluate based on job history alone.
3. AI utilization ability, design ability, and team adaptability cannot be measured by simple qualifications or years of experience.

As a result, the role required of recruitment platforms will change.

From now on, it will be more important than the number of registrants to consider how much data you can have.

The essence of AI recruitment is not automatic screening of resumes.

The key is how well you can read a candidate's potential three-dimensionally and convert it into a company's recruitment results.

The freelance engineer market is “another recruitment market”

Alongside the hiring of full-time employees, the freelance agent market is growing as a source of orders for companies.

For projects such as AI, data infrastructure, security, cloud migration, and core system renewal, companies cannot always secure the necessary human resources as full-time employees. In fact, there are an increasing number of cases where it is faster to use external experts on a project-by-project basis.

The basic structure of this market is simple.

Key point
        Key point
agent
        Key point
Key point
        Key point

From a company's perspective, an agent is a "window for externally procuring human resources that cannot be hired." From an engineer's perspective, they are also responsible for project development, contracts, billing, and negotiation of conditions.

However, if you look at it as an investment theme, it's not just about pretty things.

The agent's source of income is the difference between the order price received from the company and the fee paid to the engineer. It's called margin. In general, it is often talked about in the range of 10 to 30%, and around 20% is considered a common level. However, in reality, it varies considerably depending on the business flow of the project, contract terms, payment site, sales support, welfare benefits, and support content.

For example, if a company orders 1 million yen per month and the agent's margin is 20%, the engineer's remuneration is 800,000 yen per month.

What is more important is the depth of the commercial flow rather than the margin rate itself.

If the agent is ``end-direct,'' in which the agent receives orders directly from the client company, margins are relatively easy to see. If a system integrator, development company, or another intermediary company comes in between, a cut will be made at each stage, and the engineer's take-home pay will likely go down.

Key point
Key point→ agent → Key point

Key point
Key point→ Key point→ Key point→ agent → Key point

In the AI era, this multi-commercial system will come under pressure. If many of the tasks such as organizing project information, checking skill sheets, recommending candidates, and arranging interviews can be made more efficient with AI, it will be difficult to explain the high margins of intermediaries that simply transfer projects.

There are some operators such as PE-BANK that clearly specify a margin rate of 8 to 12%, and the transparency of the return rate is a differentiating factor. On the other hand, agents with private margins and deep commercial channels are likely to be caught out by both AI matching and low-margin platforms.

This is where the market temperature can be volatile. Agents are strong when there is a shortage of engineers, and when the number of projects slows down, price negotiations become tougher. What investors should look at is not only the sales scale, but also the direct contract ratio, margin rate, continuous operation rate, payment site, and depth of specialty areas.

Illustration: Freelance agent business flow and margin

Diagram: Where is the value of recruitment platforms shifting?

Market size: Human resources business is worth 10 trillion yen, but the content will change

The human resources business market is large.

According to a 2025 study by Yano Research Institute, the market size of the three major human resources related industries, which includes temporary staffing, white-collar recruitment, and re-employment support, is expected to be 9,796.2 billion yen in 2024 and 10,095.5 billion yen in 2025.

This sense of scale is important.

However, from an investor's point of view, it is difficult to think that the entire company will grow because the market is 10 trillion yen.

While the sales scale of temporary staffing is large, profit margins are difficult to achieve. Although the market size of human resource introduction and direct recruiting is small, it is easy to increase profitability with data and matching quality. AI will change the business processes themselves in the RPO, recruitment management, assessment, and engagement areas.

This is where the IT freelance agent market overlaps. In terms of size alone, the temporary staffing market is larger. However, in AI/DX projects, unit price, specialization, commercial distribution, and continuous operation rate are likely to be effective. Rather than total sales, profit is determined by how much margin you can make and how low costs you can match.

The market is large. The question is which part can make a profit.

Featured Company 1: Recruit HD (6098)

Recruit HD's strength is not only in the domestic human resources business.

We have a global HR technology base that includes Indeed and Glassdoor, and have huge data points related to recruitment, searches, applications, company reviews, and advertising operations.

In the full-year financial results for the fiscal year ending March 2026, sales were 3,697.351 billion yen and operating income was 630.567 billion yen. Its profit scale and cash generation ability are exceptional among domestic HR stocks.

However, it is a little difficult to see from here.

Recruit is close to the favorite for AI recruitment infrastructure, but the market has already priced in a fairly high level of maturity. Unless we see a recovery in Indeed's efficiency, matching accuracy, ad unit price, and demand for overseas recruitment, simply being "strong in AI" may be a weak stock price factor.

The KPI you should look at is not just the number of job advertisements posted.

• Matching accuracy
• Application rate
• Hiring decision rate
• Corporate advertising investment efficiency
• Operation cost reduction with AI
• Indeed/Glassdoor revenue recovery

It's around here.

Even in the side job/freelance field, Recruit has strong contacts with companies and job seekers. However, the profit model is different from that of a full-time freelance agent. The thinner the boundaries between recruitment advertising, scouting, introductions, and outsourcing, the more effective the recruiting data infrastructure will be, but price competition with low-margin platforms is unavoidable.

Featured Company 2: Persol HD (2181)

Persol HD is a comprehensive human resources company with a wide range of services including temporary staffing, recruitment, BPO, outsourcing, and DX support.

In the full-year financial results for the fiscal year ending March 2026, sales were 1,555,833 million yen, and operating income was 66,512 billion yen. Although sales are large and operating income is increasing, the operating profit margin structure is quite different from that of Recruit and Visional.

Persol's strength is its ability to get into the field.

In addition to recruitment, we can also combine dispatching, outsourcing, BPO, reskilling, and DX support. As the introduction of AI progresses, companies will not only have to worry about "hiring people" but also "how to rearrange their operations."

Persol can propose both human resources supply and business design here.

It is also compatible with freelance use. As companies move forward with the introduction of AI, they will need to create a system that combines full-time employees, temporary staffing, outsourcing, and BPO. Comprehensive human resources companies like Persol can handle not only one-off matching but also on-site operations.

However, from an investor's perspective, profit margin is more important than sales scale.

While temporary staffing and BPO can increase volume, personnel costs and operational burdens tend to be heavy. To what extent can AI make the back office more efficient and increase the unit cost of recruitment, placement, and training? This is the condition for re-evaluation.

Featured Company 3: Visional (4194)

Visional has a strong position in direct recruiting for high-class and specialized positions based on BizReach.

In the second quarter results for the fiscal year ending July 2026, sales were 46.610 billion yen and operating income was 12.768 billion yen. Both sales and profit growth are high, and operating income is strong.

What makes Visional interesting in the age of AI is that it is not just a recruitment media, but is close to the hiring decision-making process of companies.

Recruiting management personnel, DX personnel, AI personnel, and specialists cannot be achieved by simply posting a job posting and waiting. Companies need to actively search for, scout, woo, and even hire candidates.

This is where data quality comes into play.

The candidate's work history, skills, desire to change jobs, annual salary range, and employment conditions from the company. The more accurately these can be handled, the greater the value of AI matching will be.

The challenge with vision is high expectations.

Companies with high growth and high profits tend to assume strong market growth. Stock prices tend to react sensitively when the hiring market slows down, companies tighten their hiring budgets, and scout response rates drop.

In the context of outsourcing and side job human resources, Visional's strength lies in its ``high-priced professional human resource data''. If there is a growing trend to use external professional human resources for each management issue, in addition to hiring full-time employees, the concept of direct recruiting will also spread to freelance procurement. However, even here, the key is not the number of registrants, but rather how quickly and accurately we can bring out the human resources that meet the company's challenges.

Featured Company 4: Wantedly (3991)

Wantedly is a recruitment platform that focuses on empathy with a company's mission and culture rather than on requirements.

On the official website, Wantedly is said to be a business SNS used by over 4 million users and over 40,000 companies.

This place is unique.

Skill matching alone is not enough when recruiting in the AI ​​era. AI can read resumes, but whether candidates can really work at the company, whether their values ​​match, and whether they fit into the culture are different questions.

Wantedly has strengths in the areas of "empathy," "points of contact," and "recruitment PR."

However, it is necessary to take a calm look at the temperature of the latest financial results.

In the second quarter of the fiscal year ending August 2026, sales were 2.332 billion yen, -5.8% year-on-year, and operating income was 562 million yen, -39.1%. Although the profit level remains, the company needs sales to accelerate again in order to be viewed as a growth stock.

In order for Wantedly to be reevaluated in the age of AI recruitment, the focus will be on whether it can evolve into a data platform that connects candidates' skills, values, career history, and company contacts, rather than just a recruitment PR medium.

Points to look at major freelance agents

Not only listed companies, but also IT freelance agencies such as Levatech, Geeks, and PE-BANK are subject to comparison in this theme.

Levatec is well known for its IT and web-related freelance projects, and its strengths are its ability to acquire projects for companies and career support. Geeks' core business is IT freelance support, and it is easy to see the number of working engineers, unit price, and retention rate. PE-BANK is characterized by disclosing margin rates of 8-12%, and transparency itself is what differentiates them.

The metrics to look at in this area are slightly different from those for regular recruitment.

As matching operations become cheaper with AI, agents are asked, "How can we make that margin?" We have projects, are directly connected to companies, can perform technical evaluations, and can even handle problems after operation. If it's up to this point, it's still worth it.

The closer it comes to being a mere project bulletin board, the more price competition there is.

Four movements towards 2027

1. Transition to AI recruitment infrastructure

The value of a recruitment platform can no longer be measured solely by the number of registrants or job openings.

What companies want is results, not numbers of applications.

To what extent can AI read a candidate's potential and convert it into company retention, contribution, and productivity? This is where matching quality comes into play.

2. Low margin for freelance agents

AI will automate some parts of skill sheet confirmation, project recommendations, interview arrangements, and contract renewals.

If that happens, models that take a uniform margin of 20% or more will be held accountable. Of course, not everything will be low margin. Agents with advanced technical evaluation, direct end-to-end work, post-operation support, and troubleshooting retain their value.

What is at risk is the middle class, who have deep commercial channels, opaque margins, and are simply passing on project information. This is where AI and low-margin platforms are likely to take over.

3. Growth of specialized platform

AI, semiconductors, robotics, security, biotechnology, data infrastructure, etc. are areas that are difficult to evaluate using general-purpose recruitment media.

Even if you write "Python" or "AWS" on your resume, that alone doesn't tell you your ability.

This is the reason why specialized services and direct recruiting are likely to grow.

4. Fight against commoditization

The AI matching function itself will eventually be implemented in many services.

In other words, the inclusion of AI itself is difficult to differentiate.

The difference is that

• Quality of data held
• Ongoing contact with candidates
• Involvement in corporate recruitment operations
• Performance data after joining the company
• KPI that shows recruitment cost reduction

It is.

Platforms that do not have this capability will be involved in price competition even if they include AI functions.

KPIs that investors should look at

When looking at the AI recruitment/freelance market, sales alone are not enough.

I want to pursue the next KPI.

What is especially important is profit over sales, and reproducibility over profit.

The job market is influenced by the economy. When companies tighten their hiring budgets, recruitment and scouting fees are immediately affected.

That's why we need to look at whether the services are used even during economic downturns and how deeply they are involved in companies' recruitment operations.

Summary

The recruitment platform that wins in the AI era is not just a company that has a human resources database.

We are a company that uses AI to visualize the potential of human resources and connect it to the company's recruitment results, retention, business productivity, and the performance results of external human resources.

Recruit HD has the strengths of global HR data and Indeed/Glassdoor, Persol HD has the strengths of general human resources and business operations, Visional has the strengths of direct recruiting for high-class and specialized positions, and Wantedly has the strengths of empathic recruitment and contact points with potential people.

In the freelance field, specialized agencies such as Levatec, Geeks, and PE-BANK are filling companies' IT talent shortages in a different arena than hiring full-time employees. However, things are not safe here either. As the matching work becomes cheaper with AI, players with deep commercial channels and opaque margins will be eliminated.

However, adopting AI is not a pipe dream.

AI capabilities will soon be on par. What makes a difference in the end is the quality of data, contact points with companies, contact points with candidates, shallowness of commercial channels, monetization, and results after recruitment and operation.

For HR tech investment in 2026-2027, the explanation "we use AI" is weak.

What the market is looking at is whether AI can lower hiring costs, improve matching accuracy, shorten the commercial flow of freelance procurement, and whether it is directly linked to companies' development results. Companies that can confirm this will be more likely to be evaluated as the next human resource infrastructure.

Source/Reference materials

• Ministry of Economy, Trade and Industry related materials, “Survey on IT human resource demand and supply”
• IPA, DX Trends 2025-Digital talent development in the AI era
• Yano Research Institute, [Conducted a survey on five human resources business-related markets by 2040 (2025)] (https://www.yano.co.jp/press-release/show/press_id/4035)
• Yano Research Institute, “Conducting a survey on the human resources business market (2025)”
• Wantedly, Wantedly official website
• PE-BANK, PE-BANK official website
• Recruit HD, "Summary of Financial Results for the Fiscal Year Ending March 2026 [IFRS] (Consolidated)", Disclosure date: 2026-05-15
• Persol HD, "Summary of Financial Results for the Fiscal Year Ending March 2026 [IFRS] (Consolidated)", Disclosure date: 2026-05-14
• Visional, “Summary of Financial Results for the Second Quarter of the Fiscal Year Ending July 2026 [Japanese Standards] (Consolidated)”, Disclosure date: 2026-03-17
• Wantedly, “Summary of Financial Results for the Second Quarter of the Fiscal Year Ending August 2026 [Japanese Standards] (Consolidated)”, Disclosure date: 2026-04-14
• Confirmation date: 2026-05-29