Futureful for Recruiters
The desk that arrives with the market already in it.
Price a requisition before you take it. See every firm already working it, by name. Submit people as a ledger of facts instead of as a score. One board for every client on your desk, built on the same labour-market index that powers the employer side of Futureful.
A working recruiter desk with live market data. No account, no form. The desk is a sample firm and is labelled as one; every market figure in it is a live read you can check.
Early access · founding desks open in waves
- jobs reach, across every indexed source
- 22.2M+
- employers
- 86,000+
- local wage areas, 393 metro and 135 nonmetropolitan
- 528
- occupations
- 1,000+
The universe this desk is built on. Postings are demand signals, not headcounts, and these are the totals we index toward rather than a live count.
The problem, stated plainly
A search almost never dies because the recruiter was bad at it.
It dies for three reasons, and every one of them is an information problem someone decided to solve with effort instead of evidence.
The req was never fillable
The client set the rate against last year, against a national figure, or against what they paid the last person. Nobody priced it against the county it would actually be filled in — so the desk burned six weeks proving a number wrong that could have been checked on day one.
Somebody else was already on it
Four firms and eleven direct employers were posting the same role in the same market, some of them for months. You found out from the candidate, at the worst possible moment, in the worst possible sentence.
The client would not move
You knew the rate was wrong. You said so. But you were arguing with an opinion, and they had a budget. Nobody moves off a budget for an opinion; they move for a dated, sourced document they can take to their own finance team.
We built the information. It is the same index the employer side runs on, pointed at the questions a recruiter has to answer instead of the ones a hiring manager has.
The desk
Three moments, and the product is built around them in that order.
Win the req. Fill the req. Get paid. Everything below sits under one of those three, because everything a desk does sits under one of those three.
Win the req
Arrive at the client already knowing what the role pays here, who else is working it, and what it would take to fill it.
Read itFill the req
One board across every client, people rendered as evidence, and a submission record nobody can rewrite.
Read itGet paid
The engagement's real fee model, a real placement row, and a funnel measured honestly enough to renew on.
Read it01 · Win the req
Walk in with the market, not with a pitch.
Everything in this register exists to change one conversation: the one where a client tells you what they are willing to pay, and you have to decide whether that is a search or a favour.
Price the req before you take it
The wage band for that role in that client's county, drawn from real posted rates rather than a salary-survey estimate, with the p25, the median and the p75 named. You see what the client is offering against the market they are actually hiring in, before you commit a desk to a search that was never fillable at that number.
The named competitive set
Every employer posting that role in that market, by name, with the staffing firms flagged, how long each has been hiring for it and how recently they reposted. Not a category, not an index — the list, with the dates.
Agencies versus direct
How much of this role in this market is already held by staffing firms, which firms hold it, and the measured spread against direct employers. Posted rates only. We will never publish a bill rate or a markup we did not observe, and where the disclosure rate is thin we print the denominator beside the number instead of hiding it.
The one-page exhibit
A dated, sourced, print-first document you hand across the client's table unaltered: the band, the competing set, who moved and when, with numbered source footnotes carrying their vintages and the timestamp it was generated. The argument that moves a client on pay is evidence they can check, not a recruiter's opinion about the market.
The ramp calendar
Which month this work is actually posted in that county, so you open the search before the peak instead of during it — and so you can tell a client why their req has been open since March.
02 · Fill the req
A board your client can read, and a record they cannot argue with.
The search itself, across every client on the desk, with the two things a recruiter is actually paid for protected by construction: the quality of the evidence, and the ownership of the submission.
One board, every client
Client, req, stage. Briefed, searching, submitting, interviewing, offer out, placed. Client, stage, search and sort all live in the URL, so the board you are looking at is a link you can send to the client who signs — no export, no screenshot, no seat for them to buy.
An evidence ledger, never a score
Every person renders as a list of decomposable facts with their counts and their dates: what they have done, for how long, what they hold, what they said they want. No number on that card predicts the human being it describes, and none ever will. Role fit, below, is the one number on this platform — and it measures YOUR requirements, not them. That is what makes a submission defensible when a client pushes back on it.
Submission provenance
Who submitted whom, to which client, on what date, and what the client did next — a ledger, not a claim. The oldest and ugliest fight in this industry is ownership, and it is settled by a timestamp nobody can edit.
Interview kit, drafted from real facts
A role-tailored interview guide built from that candidate's own record and saved back to it, so the client hears the same evidence you did. Adults only, never a minor.
Your bench, and redeployment
The people you have already placed and already know, kept live against every open req on your board, so a contract ending is a redeployment rather than a cold start.
Reach that widens and never narrows
You can widen a search radius. You cannot narrow one, because a geographic filter pointed at a person is a proxy, and we will not ship one. Every widening is recorded, so a search that reached further can prove it.
03 · Get paid
The economics of the desk, in the desk.
A fee is a fact about an engagement, a placement is a row with a date and a wage on it, and a dashboard that rounds a number it could not read down to zero is worse than no dashboard.
Fee model per engagement
Contingency, retained, container or flat — recorded per client, per req, as the engagement actually is, never assumed from a default. Your economics differ by desk and the desk should not have to remember them.
A real placement row
Marking a placement writes the start date and the wage. Everything downstream — your fill measurement, your client reporting, your renewal conversation — reads that row, not a status somebody remembered to change.
Your funnel, measured honestly
Interest, week-over-week movement, the oldest wait in days, per-stage conversion and median days to first action, computed from your board and nobody else's. A count we could not read renders as unavailable, never as a zero. A zero you did not measure is the single most expensive lie an operating dashboard can tell.
The Monday briefNot built
Not built for the desk yet, and this card says so rather than selling it. The weekly brief ships today on the employer side, off a scheduler that already runs every Monday, and the recruiter version is that same machine pointed at your board. When it lands it is one scheduled email a week on what moved across every client, deduplicated so a quiet week does not mail you and no person is ever mailed twice, reading the same figures the console shows you.
Role fit
Paste the req. See who already raised a hand, ranked.
Drop in a job description and we read the requirements out of it. You adjust the rubric, and every candidate who applied to, saved or was shortlisted for your role is scored against what you actually asked for.
1 · Paste the req
Certifications, years, shift, county, posted rate. We pull them out and show you exactly what we read, in your words.
2 · Own the rubric
Change any weight, mark anything required, delete a line we should not have picked up. The score runs on the list you can see, never on the raw text.
3 · Open the arithmetic
Every score opens into the line items that produced it, including the misses. Nothing is weighted where you cannot see it.
4 · Work the list
It orders your worklist. It does not screen anyone out, and nobody leaves your pipeline because of a number.
What it will not do
It does not score people on anything you did not ask for. There is no personality estimate, no inferred trait, and no model opinion anywhere in the arithmetic. When we have not been told whether someone holds a certification, it says we were not told, which is a different sentence from they do not have it and leads you to a different next step: ask them.
If you use a ranking to screen candidates for a role in New York City, NYC Local Law 144 requires an independent bias audit and candidate notice before you rely on it. We put that in the product, not in a footnote you find later.
XrayAI
Paste the req. Meet the shortlist.
It reads the posting, not the job title.
The requisition
Tech Sales Development RepSan Diego, CA
Outbound pipeline for a Series B platform. Salesforce and Outreach daily. Two years of quota-carrying SDR work.
Terms it pulled out
- sales development representative
- Salesforce
- Outreach
- Salesloft
- quota
Every term appears in the posting. It cannot invent one.
Ranked against those requirements
- 94
Priya R.SDR, 3 yrs
Salesforce + Outreach, San Diego
- 89
Marcus T.SDR, 2 yrs
Quota carrying, Salesloft
- 85
Elena V.BDR, 2 yrs
Outreach daily, 12 mi
- 78
Dan K.Inside sales, 4 yrs
Salesforce, no SDR title
Illustrative desk. Every score carries the requirements it met.
A working recruiter desk with live market data. No account, no form. The desk is a sample firm and is labelled as one; every market figure in it is a live read you can check.
Catch-all addresses report as catch-all. A simulated push says so on the row.
Why it is honest
The constraints are the product, not the compliance page.
Every rule below is enforced by a build gate rather than by a policy document, which means it fails a deploy rather than failing a customer. There are ninety-one of those gates. These are the five a recruiter should care about.
We score the requirement. Never the person.
The line is exact and everything else here follows from it. Role fit scores a rubric YOU typed and can edit — certifications, years, shift, county, rate — and every point traces back to one of those lines. What we will not do is put a number on a human being: no match percentage, no readiness number, no prediction of how they will perform. The composite that once tried measured roughly four fifths language-model opinion, so it was deleted, and a build gate keeps it deleted across the entire codebase. A score about your own requirements is something you can defend to a client and to a regulator. A score about a person is neither.
Every number carries its basis
Every figure carries its instrument: the source it came from, the coverage date it was read at, and the n it was computed across whenever we have one. That is a type-level guarantee rather than a convention somebody can forget on a Friday, because the component takes the measurement as a required argument and will not render a number without them. A figure we could not measure shows a dash and a sentence naming what was missing, never a zero, because a zero you did not measure is indistinguishable from one you did. What you do not get is a clickable teardown of the join behind each number, and we would rather tell you where that line is than let you find it.
Real data or an honest blank
Where our index is thin for a county or a role, you get told so, in that spot, in plain language. You never get an estimate, a national average or a specialty guess wearing a local number's clothes.
Minors are absent from the query, and we name the one edge
For anyone who reaches us through the platform this is absolute and it is not an interface rule: the gate lives in the data layer, where a bug in a screen cannot open it, and a minor is never contactable, never in an export and never in a recruiter's reach. There is one boundary we would rather state than paper over. A person your recruiter finds on a public professional network arrives with no date of birth attached, so that path proves nothing about age, and we refuse to ship a check that reads like a verification when the field it would read does not exist. Those rows are tagged where they enter, so they are the exact population to re-examine the day age assurance is asked of them. A guarantee you can check the edges of is worth more than one that has none.
Contact is an operator action, and reciprocity is the employer's
Every send from this desk is an operator action, re-checked on the server at the moment it is sent rather than when the screen was drawn. What this desk does not have is the reciprocity gate, where access to a person is earned by a real interaction with your own posting, and we would rather name that than let you assume it. It is a hiring-side control and the reason is structural: reciprocity is proved against an employer's own postings, and a staffing desk carries its clients' reqs rather than postings of its own, so there is nothing here for a person to have interacted with. On the demo desk the question is moot: every person in it is generated and every contact field is reserved for fiction, so nobody real is reachable. The day this desk points at real people, the gate that protects them gets built and named here, and this card changes to say what it is.
None of this is modesty. A submission built out of dated, decomposable facts survives a client challenge; a submission built out of a ranking does not. The discipline that makes the platform honest is the same discipline that makes your work defensible.
Why this desk can exist at all
Futureful already owns both sides of the market.
The supply side
Futureful is free for people looking for work, in every state, at every age, with the protections that implies. That is not charity and it is not a loss leader — it is why we have a real relationship with the supply side of the labour market instead of renting profiles from someone who does.
The demand side
86,000+ employers and 22.2M+ jobs reach across every indexed source, with a hiring console already live on the employer side of the platform. We see demand as it is posted, not as it is reported afterwards.
And you sit between them
Which is the whole thesis. An applicant tracking system sees one company’s pipeline. A profile database sees people with no market around them. A job board sees postings with no people behind them. A recruiter needs all three at once, and that is the only kind of company that can build this.
What it replaces
Compared by category, never by logo.
We name what these tools do and what they do not, rather than naming companies, because a comparison you cannot check is an advertisement.
| A profile database | An applicant tracking system | A job board | Futureful for Recruiters | |
|---|---|---|---|---|
| What the role pays in this county | No | No | No | Yes, from posted rates |
| Who else is working this req, by name | No | No | No | Yes, with dates |
| How much of the role staffing firms hold | No | No | No | Yes, measured |
| A candidate as decomposable evidence | A profile | A résumé | An application | A dated ledger of facts |
| A ranking or score on a person | Yes | Often | Sometimes | Never, by build gate |
| Works across all your clients at once | Yes | One company | No | Yes, one board |
| Submission ownership record | No | Per company | No | Yes, timestamped |
| Where the people come from | Licensed profiles | Your inbound | Your ad spend | A platform they chose to be on |
Category descriptions, not claims about any named product. Where a category varies by vendor we say so in the cell rather than picking the least flattering example.
How a desk opens
Four steps, and the third one is the point.
You should be able to take something to a client before you have finished deciding whether you like the software.
Tell us your desk
Sectors, geographies, the clients you serve and the reqs you are carrying right now.
We stand up your board
Your clients, your reqs, your market — configured for the desk you actually run, not a generic workspace you have to bend.
Your first exhibit
One market read on a real open req of yours, so the first thing the product does for you is something you can take to a client.
You work the desk
And we measure fills, honestly, including the ones that did not happen.
Open a founding desk
Bring us one req you are stuck on.
Founding desks open in waves, and they open in the order firms ask. Tell us the desk you run and the reqs you are carrying, and the first thing we will do is run the market read on one of them and send it to you — before any conversation about the platform, and whether or not you ever open a desk here.
- A dated market read on one live req of yours, yours to keep.
- A real reply from a real address within one business day.
- No sequence, no drip, no list, and no seat you have to buy to find out what this is.
Seat pricing is set with the first wave of founding desks. We are not publishing a number we have not decided.