Two years after AI pronounced a "death sentence" on them, recruitment stocks have surprisingly doubled across the board.
The biggest winner in AI applications this year has surprisingly turned out to be recruitment stocks.
Among the 13 global publicly listed recruitment service companies tracked by Meridian Capital, 8 are listed in the US. As of the end of August, these 8 companies have recorded an average year-to-date increase of around 85%, with the median gain approaching 94%, almost doubling collectively.
Over the past two years, the downward cycle coupled with AI's substitution of recruitment personnel once made recruitment companies the most shunned entities in the capital market. From 2024 to 2025, these 8 recruitment stocks posted an average cumulative decline of roughly 48%.
Of course, this market rally is inseparable from the contribution of the cycle — the US employment cycle is rebounding, and recruitment companies are inherently high-leverage businesses. Once revenue reverses, the elasticity of profits and stock prices becomes extremely dramatic.
However, the cycle can only explain half of the situation. The other half comes precisely from AI.
On the one hand, AI automates a large number of standardized tasks and drives down delivery costs; on the other hand, AI makes public resumes increasingly unreliable, making data verified by real recruitment results even scarcer.
In other words, AI not only reduces the value of "human labor", but also raises the value of "efficiency" and "real data".
This year, we at Silicon-Based Observer will discuss exactly how recruitment stocks have staged a "comeback" in the AI era.
In the AI Era, Recruitment Companies Are Becoming More Valuable Instead
Many people may not be aware that traditional human resource recruitment is a typical business that relies heavily on expanding headcount. The more clients there are and the more positions there are, the more personnel are needed for screening, communication, verification, and follow-up.
When recruiting for a position, hundreds of resumes are received. Recruitment consultants first conduct a round of screening, then contact candidates one by one, confirm their willingness, ask about their basic information, and arrange interviews. These repetitive tasks could only be done by humans in the past.
Now AI has completely slashed the labor costs of recruitment companies overnight.
Adecco has taken a very aggressive approach this year, directly handing over standardized links such as candidate pre-screening, talent pool management, interview arrangement, and onboarding to AI Agents.
As of June this year, the company's AI Agents have completed more than 1.2 million candidate interactions in total, 250,000 of which are full AI interviews, covering around 50,000 positions.
More interestingly, over 60% of candidate interactions take place outside normal office hours.
AI does not need to clock off work.
A business that used to expand by increasing headcount can now expand with AI Agents. Adecco itself also disclosed that with this system in operation, the recruitment delivery time has been directly cut in half.
As of the second quarter, roughly 50% of Adecco's revenue is already generated from business processes with AI Agent participation or support. In the same period, the company's revenue increased by 6.6% year-on-year, while the number of employees decreased by 2%, and the sales and administrative expense ratio also dropped from 16.5% to 15.9%.
Apart from efficiency, another value of AI is reflected in data.
The reason is simple: AI makes resumes easier to write, but also harder to trust.
Current AI can already modify keywords in batches, polish experiences, and generate application materials for different positions. As a result, companies receive more and more "well-matched" resumes, but their authenticity is harder to assess.
A survey conducted by Robert Half on more than 2,000 US recruitment managers shows that 65% of respondents believe AI-optimized resumes make it harder to verify candidates' skills, 67% say AI-generated applications slow down the recruitment process, and 84% of HR teams face heavier workloads as a result.
Robert Half itself once cited an example of a senior accounting position:
At first glance at the resumes, several candidates all listed similar experiences in financial reporting and monthly closing, with little apparent difference. But further follow-up questions revealed that some only participated in the process, while others were truly responsible for the entire process from start to finish.
At this point, the value of data accumulated by recruitment companies over many years begins to show.
Take Robert Half as an example, it has more than 28 million talent files. When conducting matching, it not only looks at resumes and keywords, but also includes information such as whether candidates have successfully taken up positions in the past, their work performance, historical interactions with recruitment consultants, as well as interview and pre-onboarding verification results.
With the two factors combined, the value of AI to the recruitment industry becomes prominent.
On the one hand, AI automates a large number of standardized tasks and drives down delivery costs; on the other hand, AI makes public resumes increasingly unreliable, making data verified by real recruitment results even scarcer.
While Recruitment Stocks Stage a Collective Reversal, Why Is BOSS Zhipin an Exception?
While US recruitment stocks almost doubled collectively, BOSS Zhipin, the representative of Chinese recruitment stocks, has seen a far more muted stock price performance.
BOSS Zhipin was traded at around HK$78.5 at the end of 2025 on the Hong Kong Stock Exchange, and closed at HK$71 at the end of August this year, still down nearly 10% year-to-date.
After the financial report was released on August 25, the stock price rebounded by more than 20% in two days. But compared with the doubling rally of Manpower, Kelly and TrueBlue, it is not in the same league.
Why is there such a difference?
The first reason lies in the cycle.
Traditional human resource service giants such as Manpower and Adecco derive a large amount of their revenue from temporary staffing and flexible staffing.
Take Manpower as an example, temporary and contract staffing contributed 88.4% of its revenue in 2025, while permanent position recruitment only accounted for 2.5%.
This revenue structure determines that they are extremely sensitive to the macro employment cycle.
And the temporary staffing sector in the US has rebounded first.
Data from the US Department of Labor is very telling: the number of employees in temporary help services has increased all the way from 2.47 million in January to 2.505 million in July. On the other hand, the number of people in full-time jobs in the US has dropped from 134.4 million to 133.6 million, a decrease of nearly 900,000 people.
Recruitment companies themselves are also an industry with particularly high operating leverage, so once the revenue inflection point appears, the elasticity of profits and stock prices will be very dramatic.
AMN Healthcare shows this even more extremely. In the second quarter of this year, AMN's revenue only increased by 2% year-on-year, but adjusted EBITDA rose by 26%, and adjusted EPS directly increased by 158%.
Apart from the cycle, another difference lies in the monetization capability of AI.
In the second quarter of this year, the number of positions posted on the US online recruitment platform Indeed decreased by around 4% year-on-year. But its US business revenue grew by 30%, reaching a record high of 1.64 billion US dollars.
The number of positions decreased by 4%, but revenue rose by 30%. What filled the gap in between?
It is ARPJ — average revenue per job, which increased by 35%.
The company itself also disclosed that the most important driving force behind the surge in ARPJ is that AI has created higher value, making each customer willing to spend more.
The representative product is Premium Sponsored Jobs. At the end of 2024, Indeed began testing this product, and officially launched it to the market in 2025. Compared with the Standard package that mainly sells exposure, the Premium version has a higher price, but it directly uses AI to screen out matching candidates, proactively invite applications, and provide higher display positions. Hundreds of thousands of employers are already using this product.
In comparison, BOSS Zhipin also has many layouts in AI, but the relevant revenue is still at a relatively low level. In the first quarter, BOSS Zhipin's AI facilitated closed-loop service revenue was about 50 million RMB, and the overall revenue scale is still relatively small.
Summary
Looking back, what AI has truly changed is the value distribution across the entire recruitment industry.
Standardized processes such as resume screening, candidate ranking, and interview scheduling are rapidly depreciating. In contrast, links highly tied to results, such as verifying authenticity and improving matching success rates, are becoming scarcer.
The value of the recruitment industry is shifting from "processing information" to "being accountable for results". The closer one is to the final result, the more valuable one becomes.
This article is from the WeChat Official Account "Silicon-Based Observer Pro", written by Yuanyuan, and published with authorization from 36Kr.