Last week, eMarketer published a piece with a headline that would make any recruitment agency owner's stomach tighten: “Recruiting shifts from agencies to algorithms as tech platforms move in.” The article cited LinkedIn's new AI Hiring Assistant — due for global rollout — and an OpenAI Jobs Platform launching mid-2026, both positioned to absorb the sourcing, screening, and scheduling work that agencies have traditionally charged for. Bloomberg's accompanying analysis was blunter: revenue pressure will force agencies to cut recruiters to manage costs.

The panicked response — and there has been some of it across LinkedIn and recruitment forums this week — is entirely understandable. If employers can screen hundreds of CVs in seconds, rank candidates, run first-round interviews, and manage scheduling workflows inside LinkedIn's own ecosystem, why would they pay an agency a fee worth 20% of first-year salary for the privilege?

But here is the thing. We have been here before. Not in recruitment — in travel.

And what happened in travel should give every thoughtful agency owner not fear, but clarity.

The Travel Agent Paradox

When Expedia launched in 1996, the consensus was swift: travel agents were finished. Within a decade, anyone with an internet connection could search flights, compare hotels, build an itinerary, and book the whole thing without ever speaking to a human. The tools were genuinely good. They were faster, cheaper, and available at midnight from a sofa.

The number of travel agents in the UK fell sharply. Tens of thousands of jobs disappeared. High-street agencies that had thrived for decades closed their doors. The Expedia thesis appeared correct: an intelligent platform had made the intermediary redundant.

And yet.

By 2019 — two full decades into the internet travel revolution — the luxury travel agency sector was growing faster than it had in years. Specialist operators posting record revenues. Independent travel consultants charging retainer fees. A new generation of high-end travel agencies that looked nothing like the old ones: no high-street shops, no package holiday brochures, no script-reading clerks. Instead: genuine experts with deep knowledge of specific regions, relationships with properties that would never appear on Booking.com, and the ability to rescue a honeymoon when the monsoon arrives three weeks early.

What Expedia killed was the generalist, low-value middle layer. The agent who was essentially a booking interface with a phone line. The agent who added nothing to the transaction except a mild administrative convenience.

What Expedia could not touch — and what LinkedIn's AI Hiring Assistant cannot touch either — was expertise, relationships, and the ability to handle genuinely complex, high-stakes situations that algorithms systematically mishandle.

The Travel Agent Paradox is this: platforms kill the average and crown the excellent. They don't eliminate the profession. They eliminate the excuse to be ordinary.

What LinkedIn's AI Can Actually Do — and Where It Breaks Down

Let's be precise about the threat, because vague fear is less useful than a clear-eyed assessment.

LinkedIn's Hiring Assistant is, by all accounts, genuinely impressive at the bottom of the funnel. It can parse a job description, generate a ranked list of profiles from LinkedIn's 900-million-user database, send initial outreach at scale, filter responses, and manage interview scheduling — all without a recruiter touching the process. For high-volume roles with well-defined requirements and a large supply of qualified candidates, this is a real capability that will save in-house talent teams significant time and money.

The 52% of companies already reporting reduced recruitment agency dependency (Second Talent, 2026) are mostly talking about exactly this kind of hiring: software engineers, sales development representatives, junior finance roles, operations coordinators. High volume. Standardised requirements. Easy to screen algorithmically. Easy to source from a large pool.

But recruitment is not homogeneous. And the roles where agencies earn their largest fees — and where their clients feel genuine pain when the hire goes wrong — are precisely the roles that algorithmic platforms handle worst.

Consider what LinkedIn's AI cannot do:

It cannot read the subtext of a conversation with a passive candidate who has been at the same firm for nine years and is quietly ready to move, but only for something specific. It cannot interpret the brief from a CEO who says she wants “someone entrepreneurial” but actually means someone who will not challenge her authority. It cannot weigh the cultural tension between a candidate's management style and the team they would be walking into. It cannot sense that the role as written is not really the role that needs filling, and say so.

It cannot maintain a relationship with a CFO candidate over eighteen months, staying in contact without being transactional, so that when the right opportunity appears you are the first call she makes. It cannot handle the conversation where a shortlisted candidate gets a counter-offer the night before the final round. It cannot tell a client, clearly and respectfully, that their interview process is damaging their employer brand and costing them candidates.

These are not edge cases. For senior, specialist, and executive roles, they are the job. And the agencies who have been building practices around these capabilities — rather than around the administrative work that LinkedIn is now automating away — are not in trouble. They are in an increasingly strong position.

The Agencies That Will Not Survive

This is where I want to be honest, because the Travel Agent Paradox cuts both ways.

The travel agents who were destroyed by Expedia were not destroyed because they worked hard or because they cared about their clients. Many of them did both. They were destroyed because their value proposition — “I can book this trip for you” — ceased to be meaningful once a better tool existed for booking trips.

In recruitment, the equivalent agencies are those whose value proposition is essentially: “We can find you some CVs.” If that is what you are selling, LinkedIn's AI Hiring Assistant is not coming for your market in 2026. It has already been coming for it since the first LinkedIn InMail was sent in 2003. The platform has simply now automated the final mile.

The specific practices most at risk are the ones that have always been most at risk:

Multi-agency contingency on easy-to-fill roles. If a client is briefing you alongside four other agencies on a role that appears on every job board, the margin was already thin and the relationship was already transactional. An AI platform that replaces this entire workflow is not disrupting a healthy business model — it is ending a dysfunctional one.

CV-forwarding without insight. Sending a client six CVs with a cover note that says “please find attached” is a service that has been dying for years. Clients do not want a PDF pipeline. They want judgment. As we covered in The Waiting Room Test, the agencies that win in 2026 are the ones that create structured, accountable feedback loops — not those that post CVs into an inbox and wait.

Volume over depth. Working forty roles simultaneously, touching each one briefly, closing the easy ones, and abandoning the hard ones when the deadline approaches — this model depends on being the fastest mover in a field where speed was the primary differentiator. AI platforms are faster than any recruiter will ever be at the sourcing layer. This is simply a fact.

The Agencies That Will Thrive

Back to the travel agent analogy. The operators who thrived post-Expedia shared several characteristics that map almost exactly onto what distinguishes strong recruitment agencies in 2026.

They moved upmarket. Luxury travel agents stopped competing on price or convenience and started competing on access and expertise. They cultivated relationships with hotels, guides, and operators that no platform could replicate. The fee for their service stopped being a cost-of-convenience and became a genuine investment in outcomes that could not be achieved any other way.

In recruitment terms, this means a deliberate move towards retained mandates, specialist niches, and executive search — where the fee is justified not by the search activity but by the quality of judgment applied. We explored this in detail in The Chauffeur Test. If you haven’t read it, now is a good moment.

They became expert, not generalist. The most successful post-disruption travel agents are deeply specialist — East Africa safari, Japan cultural travel, expedition cruising. They know things that cannot be Googled without spending hours, and they have relationships that cannot be replicated by creating a Booking.com account. Their knowledge has genuine scarcity value.

In recruitment, depth beats breadth. An agency that has been placing CFOs in Series B-to-D technology companies for eight years has something that LinkedIn's AI Hiring Assistant cannot produce: a genuine understanding of what separates the CFOs who scale successfully from those who don’t, relationships with the exact talent pool, and the institutional knowledge to run a process that generates confident decisions, not just shortlists.

They used technology to work more efficiently, not as a substitute for expertise. The surviving travel agencies did not reject the internet — they used it to manage logistics more efficiently so they could spend more time on the high-value advisory work that humans actually needed. The agency that books your flights, hotel, and transfers automatically has freed up two hours to spend understanding exactly why you want to spend three weeks in rural Japan rather than just Kyoto.

This is precisely where AI tools like Yena create real leverage for recruitment agencies. The AI handles the mechanical layer — parsing CVs, surfacing matched candidates from your database, generating first-draft outreach, tracking pipeline stages — which frees the recruiter to spend more time on the relationship and advisory work that actually builds client trust and candidate goodwill. As we explored in The Gym Membership Problem, the value shift in recruitment is moving from volume activity to intelligent interpretation.

The Presentation Problem: Where Most Agencies Are Still Leaving Money on the Table

There is one area where I see most recruitment agencies genuinely underperforming relative to what is now possible — and it directly affects how clients perceive the value of the relationship.

Think about how a top management consultancy presents its work. McKinsey does not send a client a folder of analyst CVs. It prepares a structured recommendation: this is who we think is right, this is why, this is the supporting evidence, these are the trade-offs you should understand. The client pays for the recommendation, not the research. The research is the proof of work behind the recommendation.

Most recruitment agencies still present candidates the way McKinsey’s equivalent would have presented research in 1987. Six CVs in a PDF attachment. A covering email. An expectation that the client will do the interpretation themselves.

In a world where LinkedIn's AI is about to offer employers AI-generated candidate summaries, ranked shortlists, and structured comparison views — all free, all inside the platform they already live in — a recruitment agency that presents candidates as a PDF stack is competing on the wrong dimension.

The standard needs to be: structured candidate briefs with genuine analytical content, presented in a format that makes client decision-making easy, with a mechanism for structured feedback that allows the search to iterate quickly. This is precisely the kind of McKinsey-grade presentation that Yena's Client Portal enables — a white-labelled, consultancy-grade shortlist interface that makes agencies look like top-tier advisors, not CV forwarders.

The agencies that can present this way — and back it up with genuine expertise in their niche — are not competing with LinkedIn's AI. They are operating in a different league entirely.

Three Practical Steps for 2026

If you have read this far, you are probably not the agency that is going to be blindsided by LinkedIn's AI platform. You are already thinking about differentiation. Here is what I would focus on if I were running a recruitment agency right now.

1. Audit your role mix. What percentage of your revenue comes from roles that LinkedIn's AI Hiring Assistant could genuinely handle well — high-volume, standard requirements, large candidate pool? Be honest. If it is above 60%, you have a structural question to answer in the next twelve months. If it is below 30%, you are probably already operating in the segment where human expertise still commands premium fees.

2. Rebuild your candidate database as a strategic asset. The most important differentiator for any specialist agency is the quality of its candidate relationships — people who trust you, who have spoken to you candidly about what they are looking for, whose career motivations you understand better than any platform ever will. This intelligence lives in your database, or it lives nowhere. If your database is an Excel spreadsheet (and according to industry surveys, 73% of European boutique agencies are still running at least part of their candidate tracking in Excel), that relationship capital is largely inaccessible. It cannot be searched, matched, or activated at the moment a relevant role appears. That is not a technology problem — it is a strategic one. Getting it right is now urgent, not just sensible.

3. Differentiate on insight delivery, not search activity. What clients will pay premium fees for in 2026 is not the act of finding candidates — LinkedIn's AI can find candidates. What they will pay for is the interpretation, the recommendation, and the confidence that comes from working with someone who genuinely knows the landscape. Every client touchpoint is an opportunity to demonstrate that expertise: the brief conversation, the mid-search update, the candidate presentation, the post-placement review. If you are not investing in the quality of these interactions — and in tools that make them more structured and professional — you are letting your differentiation erode by default. The transparency and structure of how you present information to clients is now as important as the quality of the search itself.

The Correct Response to an Algorithmic Threat

When Expedia launched, the worst possible response for a travel agent was to fight the platform — to insist that clients couldn’t really book a flight online, that the experience wouldn’t be as good, that the technology was a fad. The agencies that said this were gone within five years.

The second-worst response was to try to out-compete Expedia on its own terms — to become a slightly more convenient version of an online booking tool. The agencies that tried this also mostly failed, because platforms will always win on convenience at scale.

The correct response was to accept what the platform was good at, let it do that work, and put all available energy into the things it could never do: expertise, access, relationships, and the ability to handle the genuinely complex and high-stakes.

For recruitment agencies in 2026, the same logic applies. LinkedIn's AI Hiring Assistant is not your enemy. It is an accelerant. It will commoditise the generalist middle ground, which was already commoditising anyway. It will simultaneously make the specialist, high-expertise, relationship-driven end of the market more valuable — because the contrast between what an algorithm can do and what a genuinely expert recruiter can do will become starker, not blurrier.

The question is not whether AI platforms change the recruitment industry. They clearly do. The question is whether your agency is on the right side of that change — using AI to operate more efficiently, building genuine expertise that cannot be replicated by a database query, and presenting your work in a way that makes your judgment unmistakably visible to clients.

The travel agents who survived Expedia were not lucky. They were excellent at something the platform could not be. That option is still available to every recruitment agency that makes the decision to pursue it.

A Note on Using AI Yourself

One of the more useful reframes I have found for agency owners wrestling with this question: AI is not just a threat to your business model. It is also, right now, the most powerful operational advantage available to boutique agencies who want to punch above their weight.

A two-person agency with the right AI-native tooling can run a search process — from initial database matching through candidate outreach to structured client presentation — that would have required a team of five just three years ago. That is not a marginal efficiency gain. That is a structural change in what small agencies can credibly offer.

If LinkedIn’s AI Hiring Assistant makes it easier for employers to fill their volume, standardised roles in-house, the response for a boutique agency is not to fight that trend — it is to redirect capacity freed up from commodity work into the high-complexity mandates where your expertise is genuinely irreplaceable. And to use every available AI tool to make your operation efficient enough that those high-fee assignments produce margins that make the economics work.

The Travel Agent Paradox is ultimately an optimistic story — but only for those who read it correctly. Platforms kill the average. But they create space for the excellent to be excellent in ways the old market structure never fully rewarded.

The disruption is real. The opportunity is equally real. The agencies that understand this — and act on it — will look back on 2026 as the year the market cleared out the noise and left room for the professionals.