The intermediaries are the only party the system is designed around.
Recruitment is one of the few large professional services markets where the person with the most at stake has the least control. A senior professional who wants to move relies on agencies that are paid by the other side, job boards built for volume, and HR screens run by people who cannot assess the work. Employers pay 15 to 30% of first-year salary for this arrangement and keep nothing once the hire is made [6][9]. Both sides lose.
Overture starts from an observation drawn from twenty years of executive search in digital assets and financial services: the placements that work are the ones where a strong candidate reaches the decision-maker who owns a problem, with evidence of how they solve it, before any gate is involved. Everything else is friction. A platform that reproduces that path at scale, with the candidate as the client, removes the fee, the noise and the gatekeeper at once.
The model has four parts.
Evidence, not keywords
A recorded interview and CV are combined to establish what a candidate actually does.
Prediction, not search
Overture works out the realistic next move and the companies with that problem now.
Direct, anonymous approach
The candidate is introduced to named decision-makers without revealing who they are until they choose to.
Aligned economics
The candidate pays a modest subscription; nobody is paid a percentage of anyone's salary.
The broken state of modern recruitment
The agency model: high cost, mixed value
Agencies exist to solve a real problem. They reach passive talent, take overhead off internal teams, and move quickly on specialist roles [4][5]. The price of that service is out of proportion to what is retained.
- Fee structure. Contingency fees run at 15 to 30% of first-year salary, which is $24,000 to $36,000 on a $120,000 engineering hire [6][7][8].
- Hidden costs. Several agencies competing on the same role produce duplicated outreach, confused candidates and damage to the employer's brand [9].
- Misaligned incentives. Agencies are rewarded for speed and commission, not for fit. Offer acceptance runs at around 68% on agency hires against 82% for internal ones [9].
- No pipeline ownership. The talent intelligence evaporates the moment the invoice is paid. A company hiring 25 people at $100,000 average salary can spend $500,000 to $750,000 a year on fees and own nothing at the end [9].
Of first-year salary, the contingency fee on an agency hire.
[6][7][8]Offer acceptance on agency hires against internal ones.
[9]Of top talent is not looking, so a job board never reaches them.
[2]In fees for a company hiring 25 people, owning nothing at the end.
[9]The job board illusion
Job boards were meant to open hiring up. What they produced is a structural mismatch between who applies and who is wanted.
- Active-only bias. Boards reach people who are looking. More than 70% of top talent is not [2].
- Volume over fit. The platforms are optimised for applications, so hiring teams receive hundreds of keyword-tuned CVs that loosely match [2].
- Technical blind spots. In specialised domains such as Web3, quantitative trading and AI engineering, keyword matching cannot tell depth from vocabulary [2][10].
The gatekeeper problem
The most damaging failure sits at the screening stage. HR and generalist recruiters screening for PhD-level scientific work or senior engineering roles frequently have no understanding of the work itself [3]. Lacking fluency, they fall back on keyword matching, which passes unqualified candidates through and filters out exceptional ones who do not fit the template [3][10]. Technical hiring managers respond by losing trust in the whole function: irrelevant candidates, poor communication, vague job descriptions that put good people off [10].
The net effect is that the best candidates never reach the person who would hire them, and employers never see the people they most need, because an intermediary who could not assess fit stood in the way [1][3].
Why the existing fixes have not worked
The industry has tried to solve this several times. Applicant tracking systems made the gate faster, not smarter. Professional networks made passive talent visible but left the approach in the hands of recruiters, so senior people now ignore most inbound. Recruitment-as-a-service and embedded models reduce the fee but keep the same structure: an intermediary, paid by the employer, deciding who gets through [9][11].
None of these change the position of the candidate. In every model the candidate is the product, not the client. That is the design flaw. A senior professional is the one party who knows exactly what they do, what they want next and which problems they can solve. The system treats that knowledge as an input to be parsed, not as the thing that should drive the process.
A research team that answers to the candidate
Overture inverts the model. The candidate is the client. The platform's job is to establish what they actually do, work out where that is wanted, and put them in front of the decision-maker without anyone in between. Four principles govern the design.
1. Evidence, not keywords
The candidate's primary evidence is a recorded interview, not the CV. They answer structured questions in their own words; the CV supports rather than leads. This produces a record of judgement, scope and outcomes that a keyword filter cannot see and a generalist screener cannot fake. The extracted profile is shown back to the candidate for review and confirmation before anything is done with it.
2. Prediction, not search
Given the interview and CV, Overture predicts the candidate's realistic next move and the two most credible alternatives, each with the evidence that supports it. The prediction model is built on two decades of placement outcomes in digital assets, fintech and financial services: which backgrounds actually moved into which roles, and why. The candidate approves or rejects a direction. Only an approved direction goes to market.
From there Overture identifies companies with that problem now: in-house employers, specialist funds, exchanges and infrastructure firms, filtered to seniority, location and freshness. The result is a short list of organisations and the named decision-makers inside them, not a page of job adverts.
3. Direct, anonymous approach
The candidate is introduced to the decision-maker who owns the problem, framed around the problems they solve rather than the job they had. Every approach is anonymous. The employer sees the evidence and the fit; they do not see a name until the candidate decides to release it, to that employer specifically. A discreet senior professional can explore the market without the exposure that stops most of them from looking at all.
The candidate sets the targets and approves every approach. Sending is then handled by the platform, and any reply hands control straight back to the candidate. Employers that must never be approached, including current and recent ones, are excluded from the outset.
4. Aligned economics
No fee is charged on placement. The candidate pays a modest subscription for a research and outreach system that works for them. The employer pays nothing to receive a well-evidenced, pre-verified senior candidate directly. Compared with the 20 to 30% agency tax, a platform model for a company hiring three or more roles a year costs a fraction of the equivalent agency spend while retaining the pipeline and the data [9][11].
The candidate is the one party who already knows what they do and what they want next. The platform treats that as the driver of the process, not an input to be parsed.
Confirm it. Say it. Approve it. Send it.
The candidate's path has six stages. Each one ends with a decision the candidate makes; Overture never advances on their behalf.
Upload and confirm
The CV is parsed into a structured profile. The candidate corrects and confirms it.
Interview
A recorded, spoken interview around their work, decisions and results. The transcript is reviewed and approved.
Direction
Overture predicts the next move with evidence. The candidate approves one.
Targets
Companies hiring in that space now, with the decision-makers inside them. The candidate selects who to approach and who to exclude.
Approach
Anonymous, problems-I-solve outreach to the chosen decision-makers. Replies return control to the candidate.
Disclosure
The candidate releases their identity to an interested employer, one employer at a time.
Identity is verified before a candidate's first approach. This is deliberate. Digital asset and fintech markets attract impersonation and fraud, and the value of the channel to employers depends on every approach being from a real, checked person. That verification, and the evidence behind each approach, is what makes a decision-maker open the message where they would ignore a recruiter.
Why this is defensible
Domain intelligence that cannot be scraped
The prediction engine encodes twenty years of real placement outcomes in a narrow, high-value market. Which backgrounds moved into which roles, which moves failed, which titles mean what at which type of firm. That is the judgement a good specialist headhunter carries in their head, made explicit and repeatable. A generalist platform cannot reproduce it by crawling job adverts.
A trusted channel to decision-makers
Every approach is evidenced, verified and anonymous. The recipient learns quickly that a message from this channel is worth reading, which is the opposite of what they have learned about recruiter inbound. That reputation compounds with every good introduction and is very hard for a late entrant to buy.
The candidate as client
Because the candidate pays and the candidate controls every step, Overture's incentives point the same way as the user's. There is no pressure to push a placement, no split loyalty, no reason to send a CV where it should not go. That alignment is the product.
Starting narrow
The initial market is senior professionals in digital assets, fintech and adjacent financial services, in the UAE, the UK, the US and remote roles. This is a market with acute agency fees, thin domain fluency in HR functions and a high proportion of discreet, passive talent. The same architecture extends to any specialist domain where those three conditions hold.
Risks and how the design addresses them
- Fraud and impersonation. Verification is a hard gate before any approach. In the first phase it is carried out by hand; automated document and selfie verification follows funding.
- Employer response rates. Approaches go to the person who owns the problem, framed around the problem, with evidence attached. Sending volume is small and deliberate. The candidate excludes any employer that must not be approached.
- Data and consent. The candidate confirms every extracted profile, every transcript and every direction before use. Identity is never released without an explicit decision, per employer.
- Reliance on model output. Predictions carry their evidence and the candidate can reject them. Overture proposes; the candidate decides.
- Positioning. Overture is a career research and outreach system, not an automated job application tool. It will not mass-apply, and it will not act without approval.
Three models, side by side
The three models side by side on the dimensions that determine cost, quality and control.
| Dimension | Traditional agency | Job boards | Overture |
|---|---|---|---|
| Cost per hire | 15 to 30% of salary; $20K to $36K typical [6] | Low upfront, high internal time cost | Predictable subscription; ~$5K to $7K amortised [11] |
| Who is assessed | Active and some passive; surface-level vetting | Active only; keyword-matched | Passive and active seniors; interview-evidenced |
| Who decides fit | A recruiter, often without domain fluency | An applicant tracking filter | The decision-maker who owns the problem |
| Pipeline ownership | None; disappears after the placement | None | Retained by employer and candidate |
| Decision-maker access | Filtered through the recruiter | Self-serve, no curation | Direct, anonymous first approach |
| Candidate control | Low; CV sent widely | Low; applications tracked by others | Full; identity released per employer |
Figures are drawn from the published sources listed at the end of this paper and are indicative rather than audited.
The imperative for change
The current architecture of recruitment serves neither the employer nor the professional. Agencies charge a premium to add friction. Job boards produce noise. Gatekeepers without domain fluency block the very people companies most need [1][3][9]. The industry's attempts to fix this have made the same structure faster without changing who it is built for.
A platform that puts the candidate in the position of client, replaces keywords with evidence, replaces search with prediction, and connects verified senior talent directly and anonymously to the people who hire it removes the intermediary rather than re-pricing it. The theoretical case is clear and the operating model has been proven by hand for two decades. What remains is to run it at scale.
What the paper draws on
- [1]Why Recruiters Are Struggling More With Clients Than Candidates. Top Echelon. topechelon.com/recruiter-training/why-recruiters-are-struggling-more-with-clients-than-candidates/
- [2]Why Top IT Talent Rarely Comes From Job Boards. The Connors Group, May 2026. theconnorsgroup.com/2026/05/05/why-top-it-talent-rarely-comes-from-job-boards/
- [3]Why HR Can't Understand The Jobs They Recruit For. Deep Dive Recruitment. deepdiverecruitment.com/why-hr-cant-understand-the-jobs-they-recruit-for/
- [4]In-House vs Recruitment Agency: The Real Cost for U.S. CEOs. Alliance Recruitment Agency. alliancerecruitmentagency.com/in-house-vs-recruitment-agency/
- [5]Why Staffing Agencies Beat In-House Hiring in 2026. StaffNow. staffnow.uk/why-staffing-agencies-beat-in-house-hiring/
- [6]Hiring Costs: Agency vs In-House (January 2026). Dover. dover.com/blog/cost-of-hiring-agency-vs-in-house-vs-embedded-recruiting
- [7]Agency vs in-house vs embedded recruiting cost guide. Paraform. paraform.com/insights/true-cost-of-hiring-agency-vs-in-house-vs-embedded
- [8]Recruiting agency: frequently asked questions. FirstHR. firsthr.app/compare/hiring/recruiting-agency
- [9]Recruitment-as-a-Service vs Agencies vs In-House: 2025 Cost Breakdown. Alivio Search. aliviosearch.com/blog/raas-vs-agencies-cost-breakdown-2025
- [10]The Top Reasons Technical Hiring Managers Lose Trust in Recruiting. daily.dev Recruiter. recruiter.daily.dev/resources/top-reasons-technical-hiring-managers-lose-trust-recruiting/
- [11]Staffing Agency vs Recruiting Platform in 2026. The Hire Signal. thehiresignal.com/staffing/staffing-agency-vs-recruiting-platform-2026
- [12]Pros and Cons of Using a Recruitment Agency in 2026. RecruitBPM. recruitbpm.com/blog/pros-and-cons-of-using-a-recruitment-agency
- [13]Recruitment Agency: Costs, Benefits and Selection Criteria. Wide and Wise. wideandwise.co/blog/working-with-recruitment-agency-costs-benefits
- [14]The Real Cost of Running a Recruitment Agency in 2026. Execue. execue.io/blog/real-cost-running-recruitment-agency-2026-tool-stack-fees-margins
- [15]Should Companies Use Staffing Agencies in 2026? CDR General Services. cdrgeneralservices.com/should-companies-use-staffing-agencies/