An AI Recruiter Just Undercut the Agency Fee by Half

A few weeks ago the story was an AI talent agent charging the same 20 to 30% fee a human recruiter charges - matching the price exactly and pointing it at the recruiter's own clients. The natural reassurance was that an AI charging the same as a human is not really cheaper, so the human still wins on quality. That reassurance just expired.

A service called Jack & Jill runs a full contingency recruitment model powered by AI. It sources candidates, screens them, and presents them to the employer. And it charges 10% of the hire's first-year compensation - against the 15 to 30% a traditional agency charges. Same structure. Same no-win-no-fee promise: you only pay if it makes a hire. Roughly half the price. It even offers a full refund if the hire leaves or is dismissed within the first three months.

The maths a client now sees

Put concrete numbers on it. On a 200,000-dollar hire, a traditional agency at 20% invoices 40,000 dollars. Jack & Jill at 10% invoices 20,000, with a money-back guarantee for the first three months. For a company making several hires a year, that gap compounds into very real money - the kind of saving a finance team notices and asks pointed questions about.

This is no longer "an AI that costs the same as you." It is an AI that does a recognisable version of your work, charges half, and removes the client's downside risk with a refund. The client holding two invoices for the same role is now holding a 20% invoice and a 10% one, and the cheaper option comes with a guarantee.

The fee was never about the percentage

Here is the part the pricing war obscures. The recruitment fee was never really about the number. It was about who could do the work. A client paid 20% because finding and screening the right person genuinely took a skilled human weeks of effort, relationships and judgement. The price was tied to the difficulty of the task. The percentage was just the way that difficulty got expressed on an invoice.

Once an AI can do a version of that task, the logic that justified the price changes. The work is no longer scarce or hard in the same way, so the price attached to the work falls toward the new cost of doing it. Twenty becomes ten. And there is no reason to think ten is the floor - when the marginal cost of an AI making a match approaches the cost of running the software, a flat monthly subscription model is the logical next step. Several platforms are already there, swapping the placement fee entirely for a fixed subscription.

Why "we do the same thing, but better" loses

The instinct under fee pressure is to argue quality: we do what they do, but better. That argument quietly concedes the most damaging point - that you and the AI are doing the same thing. Once a client accepts that the service is fundamentally the same, "better" has to justify paying double, and better rarely clears that bar in the client's mind. A modest quality edge is not worth a 100% price premium to most buyers, especially when the cheaper option carries a refund guarantee.

If the only difference you can articulate is that you do the identical job to a higher standard, you have already lost the negotiation. The percentage becomes the whole conversation, and on percentage you cannot win against software.

Where the fee actually survives

The recruiters who hold their fee through this are not defending the percentage. They are attaching the fee to something the AI cannot put on an invoice. Three things, specifically.

First, the read on a candidate that no benchmark captures - the judgement that this person will thrive in this particular team, with this particular manager, under this particular pressure, even though their CV looks identical to three others. That is pattern recognition built from years of placements and follow-ups, not from a model trained on public profiles.

Second, the honest counsel that tells a client the role they have written is the wrong role - that the seniority is off, the salary is uncompetitive, the brief is internally contradictory, or that they should not hire at all and fix something else first. An AI optimises for filling the role as specified. A good recruiter sometimes earns their fee by refusing to.

Third, the relationship that makes you the first call - the trust and market knowledge that means a client picks up the phone to you before they post the role anywhere, because you understand their business and their history of hires. A matching engine has no relationship to be the beneficiary of.

The fee follows the value, not the task

The unifying principle is this: a fee attached to a task that has become automatable will fall to the cost of automation. A fee attached to judgement that cannot be automated holds its value regardless of what the software charges. The Jack & Jill model is not really an attack on recruiters. It is a market signal about which half of the recruiter's job has become a commodity and which half has not.

That signal is a gift if you act on it. It tells you, with unusual clarity, exactly which parts of your service to stop leading with and which parts to build your pitch around. The agencies in trouble are the ones whose entire value proposition is the half that just got cut to 10% - sourcing and screening, sold as a process. The agencies that thrive are the ones who can point a client to the judgement, the counsel and the relationship, and make the fee about that.

The question to answer before a client asks it

If a client put your fee next to a 10% AI with a refund guarantee, what is the specific reason they would still choose you? Not a vague claim about quality or experience - a specific, demonstrable reason. And the harder follow-up: is that reason actually written into how you pitch and price, or is it just quietly assumed, sitting in your head while your proposals still lead with sourcing and screening? The recruiters who survive the repricing are the ones who can answer both questions out loud, today, before a finance team forces the conversation.