AI agents
Sourcing finds candidates overnight, the interview agent records the meeting and gives an initial scorecard assessment, the calls agent makes the first call minutes after an application, the insight agent reads all the analytics and says what to do. A person confirms every result.
The agent family
Each has its own area and its own shift. What they share: the scorecard, the system's data and one rule: a person has the last word.
AI agent · Sourcing
Vacancy and brief
legal counsel AND (M&A OR contracts)
X-Ray: LinkedIn · GitHub
Under the hood
The agent reads the brief and the vacancy description and builds search queries itself: Boolean for the hh resume base, X-Ray for LinkedIn and GitHub. Queries are visible and editable; the whole team searches by one map.
Every profile is compared with the vacancy requirements. Next to the percentage you see what matched and what is missing: experience, language, salary range, location. The recruiter decides by the reasons; the number is only a hint.
Opening contacts costs money, so the agent has a daily limit and a fit threshold below which it does not open contacts. By 08:00 the digest is ready: who was found, who is already in the funnel, what it refined after your edits.
Contact limit for today
6 / 10
Digest · 08:00
AI agent · Interview
Online interview
Recording stays in the card: rewatch the candidate, reactions and answers, at any moment
After the call
You see who spoke and when: speaker cards light up as the recording plays. Key moments are pinned on the timeline in three colours: fact, check, red flag. A click on a moment seeks the recording and opens the right place in the transcript.
12:40 · Candidate: I can start in two weeks, but by contract I need to close the quarter, negotiable.
After the interview the agent gathers the results in one place. Scores for every criterion of the vacancy scorecard, hard and soft separately, and the fit percentage computed from them. A short candidate summary. Green and red flags from the conversation. Hiring risks with a plan for the offer and for onboarding. Salary expectations against the range. And motivation across seven factors, to know what will keep the person.
fit to the vacancy
A strong M&A lawyer with confirmed English and complex deal experience. Fine with office 5/2. Salary expectations above the range, start date depends on closing the quarter.
Up to five candidates side by side on the same criteria. The agent's verdict shows right on the card in the funnel. For the hiring manager a PDF profile is assembled in the company's colours: summary, scores, motivation, risks.
AI agent · Calls · pilot
Outgoing call
Dilnoza R.
Cashier · Central Mall
In detail
The screening call after an application is the main scenario, not the only one. The agent calls back those who did not pick up, reminds about the meeting the evening before and two hours ahead, offers a new time if the candidate cannot make it and asks why if they did not show up. Each scenario is switched on separately.
Nobody writes the script by hand: it is built from the vacancy questionnaire and the hiring manager's brief. The agent speaks the candidate's language, calls only during working hours, warns about recording and hands the recruiter everything it could not score itself.
Script · from the vacancy questionnaire
After the call the candidate card gets a Call tab: score, quotes, transcript, recording, what the agent decided and why. Per vacancy you see how many were reached, how many passed and how many booked a meeting.
78%
reached
61%
passed
44%
booked
3:40
avg call
AI agent · Insights
Agent sources
Open
14
TTF
31 d
CPH
$203
eNPS
+9
Funnel
48031019052Load
Asem14Dilshod4Nodira9Vacancies
Cashier47 dLegal counsel22 dMerchandiser31 dManager decisions
Legal dept6 dRetail1 dLogistics2 dOpen
14 · 3 over SLA
TTF
47 d · target 30
CPH
$71 vs $94
eNPS
+9 newcomers
questionnaire → meeting 38% · was 52%
61%43%38%Asem 14 · Dilshod 4 · imbalance ×3.5
AsemDilshodNodira3 over SLA · 7 tasks overdue
CashierSLA +17dLegal counselSLA okMerchandiserSLA +1ddecision 6 days · limit 2
Legal dept+4dRetailokLogisticsokMetrics collected · period: 30 days
In detail
The agent is not limited to one tab. It sees every number in the system: dashboard indicators, the funnel by milestones, sources and cost per hire, team load and activity, analytics of every vacancy, SLA deadlines and tasks, hiring manager decisions, retention and newcomer eNPS. The system computes, the agent reads ready metrics: candidate names and personal data never reach the model.
Dashboard
12 indicators
Funnel
6 milestones · conversions
Sources and CPH
$203 · 5 channels
Team
8 recruiters
Vacancies
14 open
SLA and tasks
3 · 7
Managers
days to decision
Retention
0–3 months · eNPS
Rules that do not depend on AI run first: a vacancy open longer than target, a gap between funnel stages, load imbalance, overdue tasks, an eNPS drop. On top of the facts AI writes short conclusions: strengths, risks, trends and actions, each linked to a metric. A conclusion is kept for a day, refreshed on demand, and your "useful" or "not" changes the next ones.
The manager does not have to walk through tabs. The agent assembles the hiring picture across the company: what is open and for how long, where SLA is broken, which tasks are overdue, which hiring managers are late with decisions, which recruiters are overloaded. It compares vacancies, recruiters, departments and managers with each other and highlights where to look first.
Hiring across the company · week 38
+2 this week
Cashier, Merchandiser, Analyst
5 with Asem
Legal dept, limit 2
Comparison
What it gives the business
What changes when sourcing, the first contact, interview breakdowns and analytics reading are done by agents under human control.
The sourcing agent searches and scores overnight, the calls agent takes the first contact, the interview agent writes the breakdown. Recruiters are busy with meetings and decisions.
Shown by:Scored profiles, calls within minutes, recruiter hours per stage
Sets the brief and parameters, reads the vacancy analytics, checks the agent's scores against their own by scorecard. Scoring is objective and explainable to the hiring manager.
Shown by:Share of interviews by scorecard, score agreement
Every agent score is per competency with quotes from the conversation; the recruiter confirms or edits. The hiring manager gets a candidate comparison and a report.
Shown by:Candidate comparison, decision time
The insight agent reads all company analytics, compares vacancies, recruiters and departments, consolidates SLA and tasks and keeps the head of recruiting informed.
Shown by:Weekly digest, deviations, SLA
Consort Group has been in recruiting since 1992. We close vacancies in this system ourselves every day, so every feature is proven on live hiring
Discuss which routine to hand to the agentsAgents on your vacancy
We give the sourcing agent your brief and show the first batch of candidates and an interview breakdown with scores
Family rules
Criteria are set in the vacancy and are the same for the recruiter, the manager and the agent.
The recruiter edits scores and confirms the result. The agent learns from that edit.
If there was no answer, the agent writes that there is no data. No guesses, no nice-sounding filler.
Vacancy, brief, scorecard, candidate history. Nothing to copy over by hand.
Three AI helpers that prepare texts for the recruiter: not agents yet, but already saving hours.
The agent starts acting on its own inside your process, and the helpers above become unlimited.
Let's start
We give the sourcing agent your brief, show the first batch of scored candidates and an interview breakdown by scorecard. Decisions stay with people.
Or write to us
The quickest reply comes to a request through the form: it reaches the team right away together with your topic.
