When Our AI Agents Look for Jobs on Our Behalf

Could this be the next big shift in recruitment? Jack & Jill, one of the emerging players in AI-powered recruitment, recently announced a $40 million Series A funding round, just 12 months after raising $20 million in seed funding.

Image generated with AI under licence.

Another wave of hype in the AI ecosystem? What is Jack & Jill actually bringing to the table?

Essentially, two AI agents. Jack represents job candidates, while Jill represents employers.

The two agents "talk" to each other, understand what both sides are looking for and try to find a match before anyone has even applied for a job.

In some ways, this reminds me of the value proposition Jobandtalent was promoting more than 15 years ago: "Don't look for a job. Let the job find you." That turned out to be a spectacular failure. More recently, casual job boards such as Jobtoday, Snagajob and CornerJob promised to help people find a job in less than 24 hours — a clear case of overpromising, then and now.

With AI now taking centre stage, Jack & Jill is putting forward a genuinely ambitious proposition: forget traditional recruitment processes and let AI do the work for us.

Jack & Jill doesn't provide much detailed data, but claims its network already includes more than 380,000 professionals, over 5,000 companies and more than 25,000 interviews generated.

Clearly, this is still a very early-stage proposition. It needs time to mature and demonstrate its value. For now, however, it has certainly caught investors' attention — which, of course, guarantees nothing.

What seems certain is that other companies will follow these pioneers, opening up a new avenue for exploring something that has long been the Holy Grail of recruitment: the perfect match.

Let's Not Get Carried Away by the Hype

Beyond the investment itself, at a valuation of approximately $158 million — which seems rather high for a company founded in 2025 that has yet to demonstrate clear traction — the real question worth asking is:

Are we beginning to move away from a job market based on searching and applying towards one where AI agents continuously represent professionals and employers?

And if we believe that hypothesis…

Does this genuinely change the model, or are we simply building a new interface on top of the same old rules?

The Big Shift: From Searching to Being Represented

For decades, the fundamental architecture of digital recruitment has barely changed. The underlying process has remained essentially the same.

Jack & Jill infographic generated with AI

A company publishes a job vacancy. Candidates find it and apply. The company receives applications, filters them, creates a shortlist and begins interviewing.

Technology has gradually been added to this basic process to extend reach, reduce costs and time, and make recruitment easier to manage: job boards, aggregators, professional networks, ATS platforms, programmatic recruitment, recommendation engines, AI-powered matching…

But fundamentally, the logic remains unchanged:

Job posting → Search → Application → Screening → Interview → Hire

Now, Jack & Jill is proposing something different.

Jack speaks with professionals to understand not only what they have done, but what they want to do next, their salary expectations, the type of company they are looking for, the conditions they would accept and what might persuade them to consider changing jobs.

Note: This is somewhat similar to what Hired (acquired by LHH Recruitment) used to do with technology professionals, or, philosophically, what a headhunter should be doing — although only some do it consistently.

From there, the two agents look for potential matches. Only when they identify sufficient mutual interest do they introduce the candidate and the company.

That means there doesn't necessarily have to be a job search, an application or even an active intention

to change jobs.

Under Jack & Jill's proposed approach, the model could evolve into something like this:

Candidate's AI Agent ↔ Employer's AI Agent

And only when both agents identify a sufficiently interesting opportunity do people become directly involved.

AI Is Creating a Problem That Needs Solving

AI is dramatically reducing the effort required to apply for a job.

Finding vacancies, tailoring a CV, preparing a cover letter and answering application questions can increasingly be automated.

The marginal cost of submitting an application is approaching zero. And that has consequences:

More applications.

Not necessarily better ones.

Companies then need more technology to filter those applications, while candidates use even more technology to get past those filters.

We end up in a rather absurd cycle:

AI to apply → AI to filter → More AI to apply → More AI to filter

The agent-to-agent approach introduces a different possibility: perhaps the answer isn't to make job applications more efficient.

Perhaps the answer is to make them unnecessary.

Instead of allowing a candidate to automatically apply for 500 vacancies, their agent could analyse thousands of opportunities and pursue only those where there is a reasonable degree of compatibility.

The employer's agent could do exactly the same with candidates.

In theory, fewer applications and better-quality matches.

But that's only in theory.

For this approach to succeed, candidates must be willing to entrust something as fundamental as finding employment to a "black box" like AI. Employers, meanwhile, must accept that the candidates selected by the system are the ones worth reviewing, potentially overlooking everyone else.

And I say this because on the candidate side, I can imagine this level of trust among certain highly qualified professionals or people who aren't in any hurry to change jobs. I have serious doubts, however, when it comes to operational roles or people who urgently need employment.

On the employer side, we've spent years providing companies with screening and automated rejection tools. Yet there's always — always — a degree of mistrust, and employers consistently want access to as many candidates as possible.

So yes, the technology can do it. But we also need a cultural shift, and that is neither easy nor quick.

The Proposition Sounds Good. Perhaps Too Good.

As I mentioned earlier, Jack & Jill is an interesting signal, but it has yet to prove that this model actually works.

It hasn't demonstrated that the approach consistently leads to hires, or that the candidates identified are better than those sourced through other channels.

For now, there's plenty of hype, but we still need meaningful performance metrics: interview-to-hire ratios, retention, candidate and employer satisfaction, quality of hire and cost per hire.

Only then can we determine whether we're genuinely looking at a superior recruitment model.

Agents Change the Experience. They Don't Eliminate the Need for a Marketplace.

Even if we replace searches and application forms with intelligent agents, we are still fundamentally operating within a two-sided marketplace.

For a candidate's agent to be useful, it needs access to a large number of relevant opportunities.

For an employer's agent to be useful, it needs access to a large number of relevant candidates.

In other words, we still need liquidity: a substantial volume of job opportunities and candidates — two of the fundamental pillars of any job board's value proposition.

The conclusion: The traditional rules of marketplaces — scale, liquidity, density and network effects — remain fully relevant.

In fact, they could become even more important.

This brings us to a statistical paradox. Let me try to explain.

We might assume that better AI-powered matching would mean needing smaller candidate databases. But the opposite could actually be true.

The better an agent becomes at understanding complex requirements, the more value it can extract from a huge pool of candidates. This is one of the underlying principles of LLMs.

For example, finding a Product Manager in Barcelona is relatively straightforward — assuming we're not being particularly selective.

But suppose we're looking for a Product Manager specialising in B2B marketplaces, with international experience, willing to work on-site three days a week, expecting a salary between €40,000 and €50,000 per year, and potentially open to changing employers within the next six months.

Suddenly, we need access to a much larger pool of candidates.

In short, the greater the level of personalisation, adaptability, screening and filtering, the larger the pool of candidates and vacancies we need to analyse.

This is a classic mistake in job boards: we get carried away adding filters, but the underlying database is too small. As a result, the percentage of searches returning zero results skyrockets, damaging the user experience.

Marketplace size and depth are essential — and even more so in an AI-agent-driven model.

The Challenge Facing Specialised AI Agent Platforms

Players such as Jack & Jill start from a particular assumption: you need a specialised agent to represent you professionally.

But… what if your career agent wasn't a dedicated application? What if it was simply your personal AI agent?

An agent that already knows your professional background, projects, skills, interests, what you're learning, your schedule, your preferences and potentially much of your personal context.

In that scenario, perhaps Jack becomes unnecessary, doesn't it? ChatGPT has already started making some early moves into employment-related services — still very basic ones — and I'm sure we'll see further developments as more AI assistants explore this space.

On the employer side, something similar could happen. The agent representing a company could easily operate within its existing ATS, HCM or the tools recruiters use every day.

We could therefore see this happening within Pandapé, Hiring Room, Microsoft, LinkedIn, SuccessFactors, Oracle, Personio or virtually any major ATS platform.

The eventual model might look something like this:

Personal AI Agent <-> Corporate Hiring Agent

Which raises another question: what genuinely differentiates these specialised agents? And does that differentiation justify their valuations and the investment they've attracted?

The key lies in who controls the data — who has access to candidates and vacancies (and, consequently, employers). In other words, something that has always been central to the traditional model.

For these new specialised agents to work, they need access to information such as:

• Candidate: Identity, experience, skills, aspirations, salary expectations, availability and career intentions.

• Employer: Vacancies, requirements, salaries, company culture, future needs and hiring history.

  
  

And, to be precise, each player holds a different piece of the picture:

• Job boards: Vacancies, candidates, historical hiring needs and job-search intent.

• LinkedIn: Professional identity and the professional network graph (connections between profiles).

• ATS/HCM platforms: Candidates, recruitment processes, positions, hires and skills development needs.

These new players want to bring all of this together. That's no easy challenge, especially in an HR Tech market increasingly moving towards ecosystems where different providers contribute complementary pieces of the puzzle.

Their ambition is to shift the value proposition away from an audience-driven model — after all, job boards are essentially specialised media platforms — towards a model built around:

Data + Access + Liquidity

But let's not forget that unless someone gives you access to that information, you ultimately still need the audience. Otherwise, you'll always be a solution dependent on third parties.

What Happens to Job Boards?

People have been predicting their demise for years, but job boards are far from dead.

What may happen, however, is that their value proposition will need to evolve.

Today, a traditional job board is a destination where candidates search for opportunities published by employers and submit applications.

The next step could be to become a major source of high-quality, structured information that machines can query.

The key performance indicator might therefore shift from:

How many millions of users visit my platform?

To:

How many queries do AI agents make against my inventory?

The interface would become less important — something we're already beginning to see on the employer side with ATS platforms — while the data layer would gain significant value.

That could bring changes to SEO and SEM strategies, programmatic recruitment, CV database management, ATS integrations, AI agent integrations and even business models based on "pay & pray" or performance-based pricing.

For job boards, the biggest transformation may be adapting to become infrastructure designed for AI agents.

The Signal

Jack & Jill is by no means the confirmed winner of this emerging model. It hasn't proved anything yet. But it is an interesting signal, beyond the investment hype.

AI agents aren't going to replace today's job boards, ATS platforms and other recruitment technologies overnight.

What is clear, however, is that AI is spreading incredibly quickly, like an oil slick, and we need to pay close attention to both the opportunities it creates and the limitations it still faces.

We now have a signal — one that actually first appeared more than 15 years ago — challenging one of the fundamental assumptions of digital recruitment: that candidates and employers need to actively search for one another.

If agents can continuously represent both sides, the market could evolve from a model built around searches and applications to one based on intent, representation and continuous matching.

But technology doesn't make the economic rules of the market disappear.

New players will still need scale, liquidity, density and inventory. And they'll have to demonstrate something considerably more difficult than generating matches or interviews: that their agents actually deliver better hires.

The real race isn't about building the best AI recruiter.

It's much more about who will be able to represent candidates and employers while securing access to the data and inventory needed to make those agents genuinely useful.

That's where the real challenge of this potential evolution lies..