RecBuzz 2026. Budapest.

The Application Flood

Job boards are worried. The number of applications per job posting has exploded over the last year, and everyone at RecBuzz seemed to agree: it is only going to get worse. Companies like Seek and Hellowork explicitly talked about the need to differentiate high-impact applications from the noise.

The proof that automatic applications are working is that job seekers are willing to pay for them. JobCopilot has achieved something rare in the industry: making the candidate pay. Candidates find the application process so annoying and time-consuming that they will pay someone (or something) to do it for them.

I arrived at the conference thinking these tools were the big enemy of job boards. I was wrong. JobCopilot offers a white-label version that job boards can install on their own platform. Job boards do this for three reasons: they provide a better experience to job seekers, they get an extra revenue stream from the candidate side (JobCopilot shares revenue), and if they do not offer it, candidates will use it anyway.

Matching as the Response

If every job ad is flooded with applications, then the application itself stops being a meaningful signal. Someone clicks apply and does not finish. Or they finish and get discarded. Or they get accepted and do not show up to the interview. When applying is nearly effortless, the act of applying tells you almost nothing.

The intuitive response is to focus on matching the job with the candidate profile. If you can match them well enough, only the relevant candidates apply. This has been tried for years, but with the new AI tools it is becoming more feasible. And the advanced matching is no longer reserved for the biggest job boards. Companies like Nobl.ai and Nejo are making this technology affordable for mid-size boards.

On top of matching, companies like Genia AI provide technology for video interviews with already-matched candidates. The AI does not make hiring decisions. It asks questions to collect extra context that helps the recruiter. For profiles that are not that complete, this technology could even be used before the matching step, to gather more information after the application and make it reusable across platforms.

But Is This the Right Approach?

At first sight, it looks like job boards are trying to move up the funnel, getting closer to the HR recruiter by offering matching and screening. That is interesting because it is a way to differentiate and provide more value to employers.

But I see potential problems. Sophisticated matching algorithms can match job offers with candidate profiles. True. But the same technology can also be used to polish CVs to fool the matching algorithms. And some candidates may feel offended by being interviewed by a machine. The candidate does not care that there are many applicants and that the hiring manager is overwhelmed. If a hiring manager does not want to spend time interviewing the candidate, why would the candidate want to spend time talking to a machine?

We might even see a near future where AI agents apply on behalf of candidates and AI agents interview them on behalf of employers. Absurd? Maybe. But the direction is there.

Things are not binary. This technology can work well for certain audiences and certain roles. For junior positions, it can be a good way to filter candidates and save time. For senior positions, it can create a bad experience and damage the company's image.

The problem was initially triggered by the massive amount of applications. Is matching the right way to use AI, or just the intuitive way that makes sense based on what job boards have been doing for the last 20 years? Short personal answer: I do not know.

What I do know is that the best recruitment experience I have ever had was when a headhunter approached me for a specific role that seemed tailored for my skills. Maybe AI can be used to deliver that experience at scale.

Before the Interview

Maybe the candidate never receives an answer. Maybe a meeting is scheduled and the candidate never attends. A significant amount of time can pass between the application and the first meeting.

Guhuza takes a different approach. You do not apply for jobs. The algorithm matches you with suitable offers before any application happens. It feels similar to job recommendations on other platforms, but the quality of the match might be significantly better. The match happens before the application, not after, making the experience feel closer to what you would get if a headhunter reached out to you.

This also makes it harder to embellish your CV for specific positions, because you do not know which jobs will match. The candidate cannot optimize for different roles. It is in their interest to have a CV that honestly reflects their skills and expectations.

Pay per Hire

Pay per hire remains the holy grail of job board pricing. But it is hard to achieve. Companies like Seek and Hellowork ask candidates who applied through their platforms if they actually got the job. They even integrate through ATS to get what is called Disposition Data.

At first sight this makes a lot of sense. But the sector complains that employers often do not keep their ATS updated, and that the ATS market is extremely fragmented, with many different systems and no real incentive for them to facilitate integration. Still, Hellowork, one of the fastest-growing job boards in France over the last years, is actively using Disposition Data.

AI Agents on Both Sides

By the end of the first conference day, Bernhard Deussner (ex-Indeed) made an observation that stuck with me. He mentioned he missed seeing AI systems that handle both the employer and the candidate side, in a way that makes the hiring process feel more natural.

The candidate has an AI agent that applies on their behalf. The employer has an AI agent that screens applications and schedules meetings. The agents communicate with each other to find the best matches and provide feedback. The candidate does not spend time applying. The employer does not spend time screening. Both focus on the actual conversation that matters.

Still, these technologies will require profile-matching algorithms, processes to enrich the context of both the job offer and the candidate CV, and anti-cheating mechanisms to prevent embellished CVs or fake job offers.

Trust and Verification

CVs can be embellished. Phone calls and video calls can be faked by AI. But there are old ways to fix these problems:

• •Including third parties with reputation, like universities or online academies.
• •A government-issued proof that the candidate was actually employed by a company you respect.
• •Tests that assess real skills, though this is harder for soft skills.
• •And maybe it is inevitable that a growing share of the workforce will become freelancers, in which case platforms like Upwork help: you can see what freelancers do, what previous employers think of them, start with low commitment, pay for results, and increase commitment as trust builds. This is also a way to skip those recruitment tests where candidates usually have to work for free to prove their value. The candidate still works, but they can leverage what previous employers said in the past, and also get paid for the initial test, which is actually productive and useful for the employer. It is true though that for some roles this is more complex.

Moving Up the Funnel

Regardless of the specific solution, the pattern is clear. Job boards are being forced to move up the recruitment funnel to survive. Posting jobs and collecting applications is no longer enough. The value is shifting toward matching, screening, verification, and ultimately: delivering hires, not clicks.

The job boards that figure this out will thrive. The ones that keep selling visibility to an audience that is quietly being automated will not.

blog.recbuzzRecap.conclusionP3

blog.recbuzzRecap.conclusionP4