Korn Ferry Paid $1.1B for AMS. The Math Is 22 Placements Per Recruiter.

On June 29, 2026, Korn Ferry announced a definitive agreement to buy AMS from OMERS Private Equity for £850M (about $1.1B), stitching together a 16,000-person recruiting operation that markets itself as placing a professional "every 90 seconds." If you are a talent leader with a multi-year RPO contract up for renewal in the next four quarters, this deal is not a press release. It is the moment your procurement, finance, and hiring managers all pull the paperwork out of the drawer and ask the same question: are we still paying the right people to do this?

The honest answer in 2026 is: probably not, and definitely not on the same terms.

What Korn Ferry actually bought

Strip the marketing off the release and the transaction is a fixed-income instrument. Korn Ferry paid £659M (~$881M) in cash and £191M (~$255M) in stock. AMS throws off roughly $650M in annual fee revenue and $100M in adjusted EBITDA today, expected to reach a $140M run-rate within a year of close. The number that matters for anyone signing an RPO contract in the next 18 months is buried lower in the release: AMS carries more than $1.5B in estimated fees remaining under existing long-term customer agreements.

That backlog is the asset. Korn Ferry is not buying recruiters or software. It is buying a locked-in book of enterprise RPO commitments and betting those commitments do not churn before AI-native sourcing changes what enterprise buyers are willing to pay for.

$1.5B

The 90-second number does not survive contact with a calculator

"A placement every 90 seconds" sounds like a productivity claim. Do the arithmetic. 90 seconds is 960 placements per day, or roughly 350,000 per year across 16,000 employees. That is about 22 placements per employee per year, which is inside the normal band for enterprise recruiting throughput. It is not exceptional. It is not a moat. It is a headline number that measures scale of contracted volume, not efficiency per recruiter.

If you are being sold on Korn Ferry-AMS scale during a renewal, ask your rep to walk you through the throughput per embedded recruiter on your account specifically. That is the only number that maps to your invoice.

The pricing gap that makes this a decision, not a debate

Enterprise RPO pricing has been remarkably sticky. Public benchmarks put full RPO at $10,000 to $50,000+ per month, with annual enterprise contracts landing between $100,000 and $500,000+. Per-hire costs typically run $3,000 to $10,000 through an RPO, or $8,000 to $15,000 per month per embedded recruiter.

Now compare that to in-house baselines. SHRM's cost-per-hire benchmark sits at $4,700. The UK CIPD median is £1,500 to £3,000. And the AI sourcing math, imperfect but directionally clear, works out to roughly $15 per hire at 10 hires per month on modern platform pricing.

You can argue with the low end of that range. You cannot argue that the gap is small.

12.6%

The global RPO market was worth $10.04B in 2025 and is projected to reach $26.1B by 2033 at a 12.6% CAGR, with some analysts pushing 17 to 20%. Large organizations already account for 57.85% of RPO revenue, and IT and telecom lead demand at 31.05%. Translation: the buyers most exposed to this decision are exactly the ones Korn Ferry-AMS is built to sell to, and exactly the ones with in-house engineering budgets large enough to run their own AI stack.

The break-even math cracks below 15 hires per year

Here is the RPO industry's own honest math, before you get to AI. At 60 hires per year on $90K salaries, a full RPO saves roughly $650K versus contingency search. That is a real win. But at 12 hires per year, two dedicated embedded recruiters at $12K per month cost $288K, which works out to $24K per hire. At that volume, contingency at $18K per hire is already cheaper with no setup, no minimums, and no multi-year commitment.

Layer AI-native sourcing on top and the break-even moves again. If one in-house sourcer with AI can produce the top-of-funnel volume of three traditional sourcers (a conservative claim in 2026), the number of annual hires at which RPO wins on pure cost pushes higher, not lower. Below the break-even, you are paying for a service model that is priced against a labor cost your competitors no longer carry.

The Korn Ferry irony: their own research already made the counter-argument

Korn Ferry's own research, cited widely in the RPO trade press, found that 69% of talent acquisition professionals reported higher-quality candidates when using AI as a sourcing tool. SHRM's 2025 Talent Trends report says 69% of HR professionals now use AI for recruiting, up from 51% a year earlier, and 89% of adopters report measurable time savings.

The category-defining case study cited on every RPO sales deck is Unilever, which used AI-driven recruitment to cut hiring costs by 50% and shrink cycle time from four months to four weeks. L'Oréal runs a chatbot named Mya for high-volume screening. AWS uses Eightfold.ai for internal mobility. Ramp and Cognition AI have closed roles up to 5x faster using AI-native sourcing.

None of those companies bought that outcome through a traditional RPO.

The buy-vs-build question in 2026 is not AI or humans. It is whose AI, on whose data, with whose lock-in.

The vendor lock-in problem the AI Act just made worse

The EU AI Act entered into force in 2024 and treats employment-related AI as high-risk. NIST's AI Risk Management Framework emphasizes governance, transparency, and monitoring. Any serious RPO renewal in 2026 now has to answer a set of questions that did not exist in the last cycle:

• Whose model is scoring your candidates?
• Where does the training data live?
• Which regulator does your RPO answer to when a candidate files a complaint?
• What happens to the candidate graph, the notes, the outreach history, and the rejection reasons when the contract ends?

RPOs bundle sourcing, screening, and analytics. That bundle means candidate data leaves your walls, and the institutional memory of who was good, who ghosted, and who almost said yes lives with the provider. When the contract ends, that memory walks out with them. In-house AI sourcing keeps that graph on your side of the wall, which is a compliance argument and an IP argument, not just a cost argument.

The counter-argument from the RPO side is legitimate: leading providers have absorbed AI as a baseline capability, not an add-on. As one trade analyst put it, AI has moved from buzzword to foundational proposition inside RPOs. Fine. Then the question becomes whether the margin they need to charge you for that AI is justified when you can run the same models yourself against your own candidate graph.

The in-house sourcer shortage is real, and it changes the answer

Here is the part most "just build in-house" arguments skip. There are only about 1,872 people in the US and UK who currently hold the title of Sourcer, Technical Sourcer, or Talent Sourcer. They cluster in the SF Bay Area, NYC, and a small set of tech-forward employers: MongoDB, Verkada, Fetch, Gartner. If you decide to walk from your RPO and rebuild a sourcing function from scratch in Q1, you are hiring from a labor pool that is smaller than a single mid-market ATS's user base.

This is the honest tension. "Build in-house" is credible in 2026 only if AI multiplies each sourcer's output by 3x or more. There are not enough humans in the market to brute-force it the old way. That is the actual reason Refolk exists: one in-house sourcer describing the person they want in plain English and getting a ranked list from GitHub, LinkedIn, and the open web replaces the specialist Boolean labor that used to justify a five-figure monthly retainer.

What the RPO renewal conversation should look like now

If you have a Korn Ferry, AMS, Cielo, Randstad, Hays, WilsonHCG, ManpowerGroup, Kelly, or Allegis contract on the desk, three questions decide the renewal.

1. What is your true annual hire volume, and where does it sit against the break-even? Under 15 hires per year, the numbers do not work at enterprise RPO pricing regardless of what AI does. Twenty to 60 hires per year is the contested zone where AI-augmented in-house teams win most of the time in 2026. Above 100 hires per year with heavy geographic spread, the RPO argument still holds, but you should be negotiating price against AI-native productivity, not against 2019 benchmarks.

2. What percentage of your hires are top-of-funnel-limited versus process-limited? RPOs are excellent at process: compliance, scheduling, coordination, onboarding paperwork. They are increasingly not the best at finding the person in the first place. If your bottleneck is "we cannot get a warm senior engineer to reply," the fix is a plain-English search across GitHub, LinkedIn, and the open web, not an embedded recruiter. Tools in that lane include SeekOut (750M+ profiles), Juicebox / PeopleGPT (800M+ profiles), hireEZ, Gem, Fetcher, AmazingHiring, and Refolk.

3. Where does the candidate graph live when you leave? This is the AI Act question in commercial clothing. Insist that any renewal include an exit clause specifying what data you get back, in what format, within what window. If the RPO cannot or will not answer, you have your answer.

The 2026-2027 window

The Korn Ferry-AMS deal is not the end of RPO. It is the consolidation move that always happens right before a category gets repriced by a new input cost. The $1.5B backlog is the incumbents' side of the bet. The 89% of AI adopters reporting time savings is the other side.

If your renewal lands in the next 12 months, use the deal as leverage. Ask for shorter terms, per-hire pricing instead of monthly retainers, hard data portability clauses, and a written answer on which model touches your candidates. Meanwhile, run a parallel pilot of AI-native sourcing on one req family (senior engineers is the usual choice) with a tool like Refolk sitting next to your ATS. Six months of that data will tell you more about the right 2027 structure than any procurement RFP will.

The Korn Ferry-AMS press release said the combined firm will "shape the future of work." The buyers on the other side of that contract get to decide whether that future includes them.

FAQ

Does the Korn Ferry AMS acquisition change my current RPO contract?

Not directly. Existing AMS contracts continue under their current terms, and Korn Ferry has explicitly told analysts that the $1.5B backlog is a revenue-visibility asset they intend to preserve. What changes is your leverage at renewal. The consolidation gives buyers a defensible reason to renegotiate scope, pricing, and data portability, and to pilot AI-native sourcing alternatives before signing another three-year commitment.

What are the real RPO alternatives in 2026 for a company hiring 20 to 60 engineers a year?

At that volume, the strongest configuration is usually one or two in-house recruiters plus an AI sourcing layer, with contingency search reserved for a small number of hard executive or specialist roles. Named tools in the AI sourcing lane include SeekOut, Juicebox, hireEZ, Gem, AmazingHiring, and Refolk. The right test is a 90-day pilot on one req family, measured against your current RPO's cost per hire and time to first qualified conversation.

Is AI recruiting actually cheaper than RPO, or is that a vendor talking point?

The direction is real, the magnitude depends on your volume. Public benchmarks put per-hire RPO cost at $3,000 to $10,000 and AI sourcing platforms closer to $15 per hire at 10 hires per month, but that AI number excludes recruiter salary, ATS cost, and internal coordination. The honest comparison at enterprise volume is closer to a 40 to 70% reduction in fully loaded cost per hire, not the 99% headline the platform math implies. Unilever's 50% reduction is a reasonable planning anchor.

How does the EU AI Act affect the RPO buy-vs-build decision?

The AI Act treats employment decisions as high-risk, which means any AI touching your candidates (whether inside your walls or your RPO's) needs documented governance, transparency, and monitoring. The practical impact on renewals is that you now need to know whose model is scoring candidates and where the data lives. Keeping sourcing in-house makes those answers easier because the candidate graph, notes, and outreach history stay on your infrastructure rather than the provider's.