Your fee has always priced two different things: the judgment your clients trust, and the hours your team burns finding and chasing people. One of those is now compressible. The other never was.
Strip any mandate to its parts and the parts sort into two piles. Your fee is justified by the first. Your cost structure is dominated by the second. Those facts coexisted for decades because pile-two work had no substitute: it was recruiter hours or nothing. That stopped being true.
Reading the real role behind the stated one. Telling a client their band is a fantasy, and being believed. Managing the wobbling candidate at offer. Standing behind the placement with a guarantee. This work compounds with every mandate done well.
Building the longlist. Outreach, the follow-up, the second follow-up. First-pass screens at seven in the evening. The calendar shuffle. The client never sees any of it, and every mandate starts this pile again from zero.
Nothing about client trust compresses. Pipeline work now does. The open question is not whether the compressible half gets compressed. It is who captures the difference: you, or the first competitor who restructures around it.
Map the published pipeline onto a mandate and the compression covers exactly the second pile. Nothing in the first pile is touched, which is the entire reason this page is addressed to you rather than at you.
On the band we cite in our published fee analysis, contingency placements run 20 to 25 percent of first-year cash compensation, per SHRM's published band; the wider 20 to 30 percent figure quoted around the industry has, as far as we could trace, no primary source. The fee buys risk transfer, the guarantee, market knowledge and accountability. None of that is threatened here. What changes is the cost of delivering against it.
A credit is spent at booking, not at sourcing. Volume beyond the published tiers, and agency-specific arrangements such as operating an account on a client's behalf, are a quoted plan, and we print no numbers for tiers we quote. Mechanics and edge cases here.
Same revenue line, lighter cost line: pipeline work priced per delivery, scaling with mandates won instead of sitting in fixed payroll waiting for them.
See the per-delivery price list → $49 to $299 a month, publishedWhen revenue only arrives on placement, every pipeline hour on a search that dies is pure loss. Cutting the cost of dead searches changes the economics of every mandate you accept, including the marginal ones you currently decline.
Your clients buy access and judgment at the top of a market; none of that compresses. Your researchers' hours do. The choice is whether the compression funds deeper client work or gets passed to a procurement department that read the same market you did.
Selling embedded capacity by the seat means selling exactly the thing being repriced. The move is repackaging ahead of the curve: sell outcomes and calibration, buy pipeline delivery on a variable line, stop renting out the hours the market is learning not to pay for.
In all three cases the sequence is identical: firms that separate judgment from pipeline in their own cost accounting get to choose their position. Firms that discover the split in a client negotiation have it chosen for them.
No pipeline product answers the phone when a search goes sideways. Clients sign with firms because of who does.
Extracting the real role from the stated role is judgment applied against a relationship, in a conversation you lead. Garbage calibration in, polite garbage out.
Counseling the counteroffer, absorbing the guarantee, carrying the placement risk. That is the mile the fee prices highest, and it stays yours.
We have not measured stage-by-stage hours on live mandates yet, so that claim stays qualitative for now, and the pilot below is designed so you never have to take it on faith. The way to test it is on your own mandate, not on our word.
What does the fee buy? Not sourcing hours, whatever the invoice implies.
It buys risk transfer: the client pays only on a hire, so the agency carries the cost of every search that dies. It buys the replacement guarantee. It buys market knowledge the client cannot assemble internally, and it buys someone to blame, which is worth more than anyone admits in procurement meetings. A client who has been burned by a bad hire is not paying for a longlist. They are paying to sleep.
None of those line items is threatened by an AI recruiter. Every one of them survives, because every one of them is relationship work.
What is threatened is the internal cost of delivering against that fee. If your competitor delivers the same placement while paying per interview-ready candidate for pipeline work, and you deliver it while paying senior recruiter hours for the same pipeline, you are running the same revenue line over a heavier cost line. That is the oldest margin story there is.
Concretely: Mira takes a role brief and returns qualified, interested candidates booked for first conversations. What lands is a person who clears the bar, wants the conversation, and is already on the calendar, with the screening evidence shipped alongside.
Two caveats belong in the record. The bar-setting is still yours: a search firm's calibration on a niche mandate is real expertise, and the loop is only as good as the brief and bar it runs against. A vague brief produces a polite, useless longlist in any system, ours included. And the delivered candidate is an input to your process, not a substitute for it. Client presentation, the offer conversation and the guarantee remain the work your fee prices, and nothing about them is on the per-delivery line.
What we deliberately do not publish is a cost-per-placement comparison, because we do not sell placements. We sell the pipeline stage. The placement, with its risk transfer and its guarantee, remains your product, sold at your fee.
The reasonable worry for a firm is brand risk: outreach in your market is outreach in your name, and a search firm's name is most of its balance sheet.
The approval mechanism exists for exactly this. Every message runs under messaging you approve before anything is sent, so the voice candidates hear is the one you chose, and tone stays a decision rather than an accident. The screening bar is written by you, in your calibration, and the evidence that ships with each candidate is material you can put in front of a client in your own presentation format.
A human delivery lead reviews every candidate before handoff, which means the floor under quality is not "whatever the model produced this week." It is a person accountable for the batch, the same structural role a search firm's own review plays before a longlist reaches the client.
For firms operating at volume, the published tiers top out at 75 delivered candidates a month; past that, the pricing page points to a quoted plan, and agency-specific arrangements belong in that conversation rather than on this page.
One operating detail worth planning for: the per-delivery line changes what a junior researcher's week looks like, and the firms that handle that deliberately will keep their best people. The hours that used to go to longlist construction and chasing become hours on client work, market mapping and calibration, which is the work researchers were hired hoping to do. The firms that instead treat the line as pure headcount reduction will discover they cut the people who held the calibration, and calibration was the input the whole pipeline runs against.
Do not restructure anything. Pick one live mandate, the kind that eats junior recruiter weeks: real bar, real market, no referral shortcut. Write the brief as you would hand it to a new researcher. Then measure the two numbers that decide the question: internal pipeline hours spent on that mandate at loaded cost, and interview-ready candidates delivered against it. If the arithmetic fails on your niche, you have lost fourteen days and learned exactly where the mechanism breaks, which is worth knowing.
Agency-operated and white-label arrangements are a quoted-plan conversation: the published plans are sold to the hiring company, and the pricing page points volume and specialized arrangements to a quote. What we publish is the per-delivery price list itself, $49 to $299 a month for 3 to 75 interview-ready candidates, which is the number a firm needs to price the pipeline stage of a mandate.
On sourcing hours, yes. On the rest of the mandate, no. Metix AI delivers qualified, interested candidates booked for first interviews. It does not manage the client relationship, run the intake, negotiate offers, or guarantee a placement, which is the work a retained fee actually prices.
A contingency fee prices the placement; per-candidate pricing prices the pipeline. On this site's published numbers, contingency fees run 20 to 25 percent of first-year cash compensation per placement, per the SHRM band our agency-fee analysis uses, while Metix AI plans run $49 to $299 a month for 3 to 75 delivered candidates. The two are not substitutes: one pays for a hire with the search risk carried by the agency, the other pays for interview-ready candidates with the hiring decision kept in-house.
One live mandate decides it: internal hours at loaded cost, against interview-ready candidates delivered.