Recruiting glossary

A careers site is the employer’s own public page where open roles are listed and candidates apply directly. It is typically the lowest-cost source of applications and the one the employer controls entirely. A good careers site is generated from the applicant tracking system so it can never show a role that was not approved.

In Elevator: the careers site is a branded site connected to the ATS, one role per requisition, with drafts never surfacing. Sourcing and talent CRM

A job board is a third-party site where employers pay to list roles and candidates search and apply. Boards vary widely in cost and in the quality of applicants they send for a given role family, which is why teams measure them per channel rather than as a group.

In Elevator: roles post to job boards, LinkedIn, agencies, the referral portal and campus from one screen, with a channel rate card per source. Sourcing and talent CRM

A talent pool is a saved group of candidates an organisation wants to keep in touch with for future roles. Pools can be static lists a recruiter curates, or living saved searches that update as records change. Examples include silver medallists from a recent loop, alumni of a campus programme who now have three years of experience, or a succession bench.

In Elevator: pools are either curated or auto-refreshing saved searches over the talent CRM. Sourcing and talent CRM

A talent CRM (candidate relationship management system) holds every person an organisation has ever met, whether or not they are currently applying, so recruiters can search, segment and nurture them. Where an ATS manages candidates through a live process, a CRM manages the relationship between processes. Serious teams run both against one candidate record.

In Elevator: the talent CRM and the ATS share a single Candidate 360 record. Sourcing and talent CRM

Candidate rediscovery is the practice of searching past applicants and known candidates for a newly opened role before sourcing externally. It is usually the cheapest and fastest source, because the organisation already holds the résumé, the assessment and the interview record.

In Elevator: reopening a requisition ranks every past candidate against it with TalentLens, and adds them to the pipeline or a pool in one action. Sourcing and talent CRM

A silver medallist is a candidate who reached the final stages of a hiring process and was not selected, but whom the team would hire for a similar role. They are the best-evidenced people in the CRM: assessed, interviewed and scored. Re-engaging them when a matching requisition opens is the classic rediscovery move.

In Elevator: an automation rule re-engages silver medallists when a matching requisition opens. Applicant tracking

A referral is a candidate introduced by a current employee, usually through a referral portal and often with a bonus payable on hire. Referrals tend to convert well and stay longer, which is why many teams fast-track them past the sourcing queue and hold them to a screening SLA.

In Elevator: referrals arrive through the referral portal and can be fast-tracked by rule, with a pool for referral inbound awaiting screen. Sourcing and talent CRM

An agency is an external recruiter paid a fee, typically a percentage of first-year compensation, when a candidate they submit is hired; retained search is the variant where a firm is paid in stages to run a senior search exclusively. Agency submissions need to be tagged at intake so fee ownership is never disputed later.

In Elevator: agency submissions are tagged for fee tracking automatically, and agency cost shows on the channel rate card. Sourcing and talent CRM

Source of hire is the channel a hired candidate originally came through, such as referral, careers site, job board, agency or rediscovery. It is the metric that tells a team which channels to fund. It is only reliable if the source is captured at the first touch and survives duplicate merges.

In Elevator: sources and channels analytics show applications, screened, interviewed, offers, hires, cost per hire and retention per source. Analytics and reporting

Cost per hire is the total spend on a channel or a process divided by the number of hires it produced. Measured per channel it shows which sources are worth their rate card. Measured across the process it is a finance figure that includes agency fees, board spend and assessment costs.

In Elevator: the Source ROI and channel spend dashboard shows cost per hire and quality per channel. Analytics and reporting

An applicant tracking system (ATS) is software that manages candidates through the stages of a hiring process, from application to offer, and keeps the record of every action taken on them. It is the system of record for a hire while the hire is in progress. Modern systems add requisitions, structured interviews, offers and analytics on top of the tracking itself.

In Elevator: the ATS runs six pipeline templates, automation rules and an append-only audit log. Applicant tracking

Résumé parsing is the automatic extraction of structured fields, such as employers, dates, skills and education, from a résumé document. Parsing is never perfect. A good parser says how confident it is per field so a recruiter fixes the doubtful ones rather than re-keying everything.

In Elevator: résumé import carries per-field confidence, and low-confidence fields are corrected in seconds. Sourcing and talent CRM

AI résumé screening is the use of a model to score or rank applicants against a role so a recruiter reviews the most relevant first. Done well, the score is explained in writing against the requisition’s own band, seniority and target date, and a written rule decides what happens at a given score rather than the model.

In Elevator: TalentLens scores six weighted competency axes with a written explanation quoting the evidence, and never uses campus signals. Screening and assessments

An asynchronous AI interview is a screening conversation a candidate completes on their own time with an AI interviewer, producing a summary a recruiter reviews later. It replaces the first recruiter phone screen for high-volume roles. The recommendation it produces is a draft for a person to confirm, not a decision.

In Elevator: AVA runs as a pipeline stage and returns a summary card with per-competency signal, integrity checks and a recommended action a recruiter confirms. Screening and assessments

An assessment battery is a set of tests assembled for a specific role, combining formats such as coding, multiple-choice, case and simulation. Batteries differ by role and grade: they test depth in a domain at a given level, not general aptitude.

In Elevator: role-specific batteries cover 12 domains in 24+ formats, from live coding to architecture whiteboard simulation. Screening and assessments

A work simulation is an assessment in which the candidate performs a realistic task from the job, such as debugging a live system or handling a customer scenario, instead of answering questions about it. It evaluates what a person does rather than what a résumé claims, and it is harder to game than a knowledge test.

In Elevator: Skill Studio provides the work simulations inside an assessment battery. Screening and assessments

Proctoring is the monitoring of a remote assessment or interview to detect impersonation, unauthorised help and other integrity breaches. It targets proxy test-takers and proxy interviewees, the two failure modes that a résumé cannot reveal.

In Elevator: assessments run on the proctoring stack that handles 12 million assessments a year and 800,000 concurrent sessions. Screening and assessments

A structured interview is one in which every candidate for a role is asked the same questions, on the same competencies, and scored on the same scale. Research consistently finds it predicts job performance better than an unstructured conversation. The structure is what makes two interviewers’ scores comparable.

In Elevator: interview loops define the rounds and competencies once, and scorecards anchor every score to a rubric. Interviews and scheduling

An interview loop is the full set of rounds a candidate goes through for a role, run as one planned sequence rather than as separate invitations. A loop names each round’s mode, duration, panel and competencies in advance. Thinking in loops rather than invites is what lets a team schedule and debrief properly.

In Elevator: a loop is attached to a pipeline stage per role and runs the same way every time. Interviews and scheduling

An interview plan is the written design of a loop: which rounds exist, who sits on each, what each round assesses and how quickly feedback is due. It is defined once per role or role family and reused, so the tenth hire is interviewed the way the first was.

In Elevator: interview plans carry rounds, mode, duration, panel roles, competencies, focus and a feedback SLA in hours. Interviews and scheduling

A scorecard is the form an interviewer completes after a round, scoring the candidate on each competency the round was meant to assess. A scorecard is only useful if the scores mean the same thing to everyone, which is what a rubric provides, and if it is submitted before the interviewer sees anyone else’s.

In Elevator: scorecards are rubric-anchored and blind until submitted; panel words go to the debrief unedited. Interviews and scheduling

A rubric is the written scale behind a scorecard, describing what each score level looks like for each competency. Without a rubric, a 3 from one interviewer and a 3 from another are different numbers. With one, they are the same judgement.

In Elevator: rubrics use four-level behavioural anchors per competency. Interviews and scheduling

A behavioural anchor is the concrete description of observable behaviour that defines one level on a rubric scale. An anchor says what the candidate did, not how the interviewer felt. It is what turns an opinion into evidence a committee can compare.

In Elevator: every scorecard competency carries four behavioural anchors. Interviews and scheduling

A debrief is the meeting after a loop where the panel reviews all scorecards together and reaches a decision. A good debrief spends its time where the panel disagrees, not on the average. The decision and its reasons are recorded so an offer can be raised against them.

In Elevator: the debrief shows every scorecard side by side on a competency heat-map with divergence surfaced, and logs the decision. Interviews and scheduling

A hiring committee is a standing group, separate from the interview panel, that reviews the evidence for a hire and approves or declines it. Committees are common for senior and leadership roles, where the cost of a wrong hire justifies a second, independent look at the scorecards.

In Elevator: the committee decision is gate two: no scorecards, no offer. Interviews and scheduling

Panel capacity is the number of interviews a team’s trained interviewers can actually give in a period, after their own work is accounted for. It is the real constraint on time to hire in most engineering organisations, and the reason self-scheduling has to be built on it rather than on a blank calendar.

In Elevator: scheduling works from a panel roster, availability grid and per-interviewer load. Interviews and scheduling

Self-scheduling lets a candidate pick an interview slot from the panel’s real availability through a link, instead of trading emails with a coordinator. It removes the longest idle gap in most processes, the wait between “we would like to interview you” and a confirmed time.

In Elevator: candidate self-schedule links are built on panel capacity and load, with calendar sync and every conflict named for a human. Interviews and scheduling

A feedback SLA is the agreed time limit, usually in hours, within which an interviewer must submit a scorecard after a round. It exists because late feedback is the most common reason a good candidate goes cold. An SLA is only real if something happens when it is missed.

In Elevator: each loop carries a feedback SLA and an automation rule chases the panel when it is overdue. Interviews and scheduling

Offer approval is the sign-off a proposed compensation package needs before it is released to a candidate. It usually involves the hiring manager and finance, and adds a compensation committee when the offer is outside the band. It is the last gate before a commitment the organisation cannot easily withdraw.

In Elevator: the offer approval inbox is visible to recruiter, hiring manager and finance at once, and band-breach routing adds an approver automatically. Offers and verified hire

A compensation band is the approved minimum-to-maximum range a role at a given grade may be offered within. Bands keep pay equitable across hires and give finance a way to approve a requisition without approving each offer individually.

In Elevator: bands are multi-currency and attached to the requisition’s grade. Offers and verified hire

A band breach is an offer whose compensation falls outside the band approved for the requisition. A breach is not a failure; it is an exception that needs an extra approver. A sensible tolerance stops a rounding difference from convening a committee.

In Elevator: a breach inserts a compensation-committee approver before final signature, with a tolerance so a 0.3% overage does not. Offers and verified hire

Recurring cost to company is the part of an offer that repeats every year, such as fixed pay and target variable, as distinct from one-time components like joining or relocation bonuses. Finance approves on the recurring figure, because it is what year two costs. The headline total is what the candidate sees.

In Elevator: the six-component compensation builder rolls up to total cost and to recurring cost separately. Offers and verified hire

A negotiation log is the record of every version of an offer, the candidate’s response to each, and who changed what. It moves the negotiation out of an inbox and onto the record, so the final signed offer can be traced back through each counter.

In Elevator: offers are versioned with a v1 to v2 diff and a counter-offer log with actor. Offers and verified hire

E-signature is the legally recognised electronic signing of an offer letter, replacing a printed and scanned document. It shortens the gap between release and acceptance and records when the letter was issued, viewed and signed.

In Elevator: letters issue from a governed template and are signed through DocuSign, Adobe Sign, Leegality or Superset e-Sign, with Aadhaar eSign via connector. Offers and verified hire

Background verification is the checking of a candidate’s identity, employment history, education, address and, where relevant, criminal and credit records before they join. It is the main defence against fabricated experience. It is only a control if joining cannot proceed until it clears.

In Elevator: seven check types run through connected vendors or Superset’s own checks, and joining stays blocked until they clear. Offers and verified hire

A verified hire is an accepted candidate whose background checks have cleared and who is therefore ready to be handed to the HRIS. The phrase marks the difference between “accepted the offer” and “safe to put on payroll”, which are not the same event.

In Elevator: required verification checks must clear before the hire record is sent to the HRIS. Offers and verified hire

A pre-hire is the record an HRIS creates for an accepted candidate before their start date, from which the employee record is built. Creating the pre-hire from the recruiting system, rather than re-keying it, is the point where most HRIS integrations earn their keep.

In Elevator: Elevator creates the pre-hire in Workday or SAP SuccessFactors with a payload preview, and the employee ID comes back to close the loop. Integrations

HRIS handoff is the transfer of a hired candidate’s data from the recruiting system to the human resources information system that owns employees. The recruiting system should show exactly what it is sending and where each field came from, and the HRIS’s own approvals should still run.

In Elevator: the handoff shows the exact payload and field provenance before it reaches the HRIS, and never edits an existing worker. Integrations

A pipeline stage is one step a candidate occupies in a hiring process, such as screen, assessment, interview, debrief, offer or verification. Stages carry an owner and a time limit, so “where is this candidate stuck and whose turn is it” is a lookup rather than a chase.

In Elevator: stage types are sourced, screen, assessment, interview, debrief, offer, verification and hired, each with an SLA in days and a named owner. Applicant tracking

An SLA (service-level agreement) in recruiting is the agreed time limit for a step, such as days a requisition may sit with an approver or hours an interviewer has to submit feedback. An SLA is a promise between teams. It only changes behaviour when a breach triggers something.

In Elevator: SLAs live on approval steps, pipeline stages and interview loops, and a breach can escalate by rule. Applicant tracking

An automation rule is a written trigger-and-action pair the system executes the same way every time, such as “assessment scored below 60, reject with the recruiter’s name”. Rules are deterministic, which makes them auditable: every outcome can be reconstructed from the rule and the log. That is why consequential decisions belong in rules, not models.

In Elevator: rules pair eight triggers with eight actions, and every rule is reconstructable from the audit log. Applicant tracking

An audit log is an append-only record of every change to a record, with who made it, when, and the value before and after. It is the evidence a compliance or fairness review runs on. A summary of it is not evidence; the rows are.

In Elevator: the audit log is append-only per requisition, candidate, interview and offer, filterable and exportable. Security and trust

Role-based access control (RBAC) grants each user the permissions of a named role, such as recruiter or hiring manager, rather than assigning permissions one by one. In recruiting it decides who sees compensation, who sees other requisitions, and who can approve. Scoping is only trustworthy if it is enforced in the query, not hidden in the interface.

In Elevator: seven roles and 43 capabilities, each full, scoped or none, enforced at the query level. Security and trust

SSO (single sign-on) lets users log in to a system with their organisation’s identity provider instead of a separate password. For an IT team it means access follows the corporate directory: when someone leaves, their access to the recruiting system ends with it.

In Elevator: SSO via SAML 2.0 and OIDC with Okta, Microsoft Entra ID, Google Workspace or any compliant provider, enforceable tenant-wide. Security and trust

SCIM (System for Cross-domain Identity Management) is the standard by which an identity provider creates, updates and deactivates user accounts in a connected application automatically. With SCIM, joiners, movers and leavers in the directory are reflected in the recruiting system without an administrator touching it.

In Elevator: SCIM 2.0 provisioning maps each directory group to a role. Security and trust

A sub-processor is a third party that a software vendor uses to process customer data on its behalf, such as a cloud host, an e-signature service or a verification vendor. Privacy law requires the vendor to list them and to tell customers when the list changes.

In Elevator: the sub-processor register is derived from the systems you actually connect, and lives in the security pack. Security and trust

A data fiduciary (the controller, in GDPR terms) decides why and how personal data is processed; a data processor processes it on the fiduciary’s instructions. Under India’s DPDP Act 2023 the employer is the data fiduciary for candidate data and the recruiting software vendor is its processor. The distinction sets who owes candidates what.

In Elevator: Superset processes candidate data on the employer’s instructions, aligned with the DPDP Act, and never uses it to train a model. Security and trust

The right to erasure is a candidate’s legal right to have their personal data deleted once there is no longer a lawful reason to keep it. A retention policy says when data should go; an erasure queue is what actually removes it and records that it did.

In Elevator: a retention schedule and a working erasure queue, because a policy without a queue is a PDF. Security and trust