Find and Interview Candidates for Every Staffing Order
The Cognitive gives a staffing desk the 2 jobs its front office leaves to recruiters: AI sourcing across ~900M public profiles with verified emails and phone numbers, and a live, two-way AI video interview that candidates book themselves, scored against one rubric per role. Your staffing ATS keeps the orders, timesheets and pay, and your recruiters spend their hours on the candidates worth submitting. A UK staffing agency on The Cognitive scaled from 200 to 1,400 interviews a month with the same 8 recruiters (see the case study at /case-studies/uk-staffing-agency-scale).
Why agencies outgrow interview-by-recruiter
- Recruiter capacity caps at ~200 screens a month; client demand doesn't.
- Rotating screeners means inconsistent evaluations, and clients notice in the shortlists.
- Recruiter hours spent on first-round screens eat the margin on every placement.
- Speed loses placements: the agency that submits scored candidates first usually wins the req.
The agency workflow on The Cognitive
- Candidates import from Bullhorn, JobDiva, JobAdder, Crelate or Recruiterflow, among 60+ ATS integrations.
- AI sourcing and your candidate pool feed each req; email sequences, from Sourcing Pro up or on a Team plan, handle first contact.
- Every candidate you invite takes a live AI interview inside the window you set, so first-round screens stop waiting on recruiter calendars.
- For candidates imported from your ATS, a note with the score, summary and report link goes back to their record. Stages never move; your recruiters move them.
- Client-ready shortlists: a read-only report link with scores, written feedback and the recording, shareable with the client as proof of screening.
What staffing agency software covers, and what it leaves to you
| Layer | Examples | The job it does | What it does not do |
|---|---|---|---|
| Front office (ATS/CRM) | Bullhorn, JobDiva, Ceipal, Avionté | Requisitions, candidate records, submissions, client relationships | Built around people who applied or are already on file; outside search and AI screening, where offered, are sold by demo or quote |
| Back and middle office | Timesheets, payroll and billing modules | Pay, bill, margin and compliance on placements already made | Nothing upstream of the placement |
| Sourcing layer | Job boards, profile databases, AI sourcing | Fills a requisition with people the database does not already hold | Says nothing about quality: a sourced list is still unscreened |
| Evaluation layer | The Cognitive | Interviews every candidate live against one rubric per req and returns evidence-scored shortlists | Not a system of record: a note with the score and report link goes back to candidates imported from your front office |
"Staffing agency software" describes 3 different products, and agencies routinely buy the first and expect the third. The front office holds requisitions and candidates, the middle office handles timesheets, pay and bill; in their core editions, neither judges whether a candidate can do the job, and where vendors add AI screening it is sold by demo or quote. That judgment is still a recruiter on a call, which is why it is the layer that caps a desk.
If you are choosing the front office itself, our ranking of staffing software for temp and contract firms compares the systems that run timesheets, payroll and redeployment, and Bullhorn pricing explains how the most common agency ATS is sold. IT desks have their own shortlist in IT staffing software.
How to source candidates for a staffing requisition
| Sourcing channel | Who it reaches | Where it falls short on an agency desk |
|---|---|---|
| Your own ATS database | Candidates you placed or submitted before | Ages fast, and the contact details rot faster than the profiles do |
| Job board posts and adverts | Active jobseekers who find the advert | Misses everyone employed and content, which on a scarce skill is most of the market |
| Manual LinkedIn boolean search | Passive candidates, one query at a time | Consultant hours per search, no rationale recorded, and nothing keeps running after the tab closes |
| Referrals from placed candidates | People inside the same networks you already reach | Narrow, and it reproduces the shape of the desk you already have |
| AI candidate sourcing | Passive candidates matched against a whole brief across ~900M profiles, with contacts enriched from 30+ sources | Still unscreened: a sourced list says nothing about whether the candidate can do the job |
Candidate sourcing is the outbound half of recruitment: searching for people who have not applied to the requisition, checking they fit it, revealing contact details, and opening the first conversation. On an agency desk it is the half that decides whether a requisition gets filled at all. The database usually holds the people who were right for the last role, not this one, and a job-board post reaches only those already looking.
The Cognitive's AI sourcing works a brief written in plain English rather than a boolean string. It searches ~900M public profiles, judges each one against the whole brief, and returns a "Why this match" line per candidate, so a match is a claim you can verify rather than a score you have to trust. Contact details are enriched from 30+ sources: a search costs 1 credit per person shown, a verified email 4 credits, a direct phone number 12, charged only on a successful reveal.
A requisition keeps its search history. Everyone the role has ever surfaced stays in a durable role pool, later searches skip the people already seen, and once your consultants have made 8 or more shortlist and pass decisions, taste memory shapes later ranking toward the profiles they shortlist. Requisitions that stay open for weeks are exactly the ones a one-off search pass fails.
Market intelligence a client will pay for at intake
Market intelligence in recruitment is what a search tells you about the market rather than about one candidate: how many people match the brief at all, where they sit, what seniority band the supply actually falls in, and which requirement is the one collapsing the pool. It is the difference between telling a client the role is hard and showing them why.
Because every search runs against a written brief, dropping or loosening one requirement re-runs the market instantly. A staffing desk can go into an intake call able to say which line in the job spec costs the most candidates, which is a consulting conversation rather than an order-taking one, and it is the conversation that changes a requisition into a fillable one before a single candidate is contacted.
Recruiting metrics that tell you whether a desk is working
| Metric | What it measures | Where it comes from |
|---|---|---|
| Time to fill | Days from requisition open to accepted offer | Your front-office ATS, per requisition |
| Time to first submittal | Hours from requisition to the first credible candidate in front of the client, the number that usually decides a shared role | Sourcing and screening speed together |
| Cost per hire | Total sourcing, advertising, tooling and desk time divided by placements made | Your own cost lines; the screening component is the one most desks never price |
| Submittal-to-interview ratio | How often a submitted candidate gets a client interview, a direct read on shortlist quality | Front office, per consultant and per client |
| Quality of hire | How placed candidates actually perform, and whether they stay past the rebate window | Needs pre-placement evidence: scored, recorded interviews make it possible to correlate at all |
| Sourcing yield | Candidates surfaced per search, reply rate, and cost per contact revealed | Sourcing tooling: 1 credit per person shown, 4 per email, 12 per phone |
Recruiting metrics are the numbers that describe how a hiring pipeline behaves: how long a requisition takes to fill, what each hire costs to make, how many candidates survive each stage, and how good the people you place turn out to be. On an agency desk they double as commercial reporting: the same numbers are what a client asks for at a quarterly review.
Only one of them is hard: quality of hire, because it needs evidence about how a candidate performed before they were placed, and a resume plus a consultant's impression is not evidence. Everything else your front-office system already records. We do not publish benchmark figures for these, because a benchmark drawn from someone else's desk tells you nothing about yours. Measure your own for a quarter and compare against that.
Agency economics
With AI Interview from $99/month taking the first-round screens, screening stops being the margin killer. Recruiters move to placements and client relationships; interview capacity scales with demand spikes instead of headcount. AI Sourcing also starts at $99/month.
Frequently Asked Questions
Does staffing agency software source candidates?
Mostly not in their core editions. Front-office staffing software such as Bullhorn, Avionté, JobDiva and Ceipal is built around candidates your agency has already met or who applied, and where a vendor adds outside search, such as Bullhorn's Amplify Search & Match, it is sold by quote. Sourcing is a separate job: The Cognitive runs a plain-English brief across ~900M public profiles, reveals verified contact details, and gives you a shortlist to interview and add to your front office.
What does sourcing cost an agency per candidate?
On The Cognitive, sourcing is charged in credits rather than seats: 1 credit for every person a search shows, 4 credits for a verified email and 12 for a direct phone number, with reveals charged only when they succeed. Sourcing credit plans start at $99/month, with per-seat Team and Business plans for a whole desk. That structure suits an agency desk because building a long-list is cheap and the cost only lands on the candidates a consultant decides to contact.
What is the best staffing agency software for candidate screening?
The Cognitive is the staffing screening layer you can open without a sales call: a live, two-way AI video interview for every candidate you invite, one rubric per req, and a scored report per candidate. Staffing platforms such as Bullhorn, Avionté and JobDiva are built for requisition and pipeline management, and the AI screening some of them list is sold by demo or quote. For candidates imported from your ATS, a note with the score, summary and report link goes back to their record, so the system you know stays.
How do staffing agencies use AI interviews with clients?
Agencies share the read-only report link as proof of screening depth: the client sees a 1 to 5 score on each criterion, the overall written feedback, the weighted score and the recording, so it is clear why each candidate was submitted.
Other solutions
BPO & Staffing Industry · Bullhorn Integration · UK Agency Case Study · Sourcing + AI Interviewing Platforms Compared · Hire by Role