Skip to main content

Answers

AI scoring and the Hiring Agent

How inploi's Hiring Agent scores applications, what it looks at, who sets the criteria, how bias is audited, and why it never rejects a candidate.

Updated

inploi is the hiring experience platform for large frontline employers. Agents do the admin. Humans do the hiring. It all runs on the ATS you already have. This page answers the questions buyers ask about inploi’s Hiring Agent, which scores applications for recruiters, and about how inploi uses AI with candidates. Each answer opens with a direct reply; the links go to the page on www.inploi.com that covers the topic. For neutral guidance on auditing AI scoring tools, see https://www.inploi.com/insight/high-volume-hiring.

How does inploi’s AI candidate scoring work?

The Hiring Agent scores every application against criteria your hiring managers set for the role category, and shows the evidence for each criterion: the line from the CV or the answer that earned the score, and why the candidate ranked where they did.

  • Criteria are set once for a category of similar roles, such as all kitchen roles or all store roles, and each criterion carries a weight in points, so what matters most for the job counts most.
  • Basic requirements are asked as screening questions and scored by fixed rules, not by a model; the weighted criteria cover the judgment calls.
  • Scores are produced as applications arrive, and recruiters sort the queue by score instead of by date applied.
  • Recruiters see the ranking and evidence in Studio and, where the ATS takes it, in the ATS.
  • Your criteria, your weightings, your sign-off.

Related: https://www.inploi.com/agents/hiring-agent

Does the AI make hiring decisions or reject candidates?

Agents rank and recommend; they never reject, and a person on your team makes every hiring decision.

  • The score changes the order of the queue, not who is seen: every application stays in the queue and can be read in full, CV included.
  • The agent marks its top matches; it never moves anyone to a new stage. A recruiter checks the top of the list, reads the evidence and decides who to interview.
  • Nothing in the product hides, holds or stops an application because of a score.

Related: https://www.inploi.com/security/ai · https://www.inploi.com/agents/hiring-agent

What does the Hiring Agent look at, and what is removed first?

Only the criteria set for the role and what the application shows against them: before any AI model sees an application, a separate step removes names, contact details and details that indicate protected characteristics.

  • The model receives the role’s criteria and a de-identified account of what the candidate has done and answered, not who they are.
  • Where a candidate answers by voice, the audio is used only to make a transcript, and the transcript is what is scored. How someone sounds is never scored.
  • It does not use information from outside the application.

Related: https://www.inploi.com/security/ai

How does inploi guard against bias in scoring, and how is it audited?

Personal details are removed before scoring, scoring uses only the criteria set for the role, every score carries readable reasoning, and inploi’s use of AI, including the Hiring Agent and the Candidate Agent, is independently audited for fairness and accuracy by Warden.

  • Independent testing with realistic test data shows whether an AI treats people differently because of who they are.
  • Whether outcomes differ across groups in your own hiring can only be measured on your own data, and inploi will work with your team on that audit.
  • A full audit trail of every agent action is available in Studio.

Related: https://www.inploi.com/security/ai · https://www.inploi.com/insight/fair-transparent-and-audited-how-inploi-is-raising-the-bar-for-ai-in-talent-acquisition

Who sets the scoring criteria, and could the criteria themselves be biased?

Your hiring managers do, with your talent team’s approval: each one starts from a template for the role category, or from a set inploi proposes from your own hiring data, then sets the weights, calibrates and submits them. Because criteria can carry bias, they should be limited to what the job genuinely requires.

  • Criteria are the role’s requirements as you define them: qualifications, availability, experience, what makes someone right for this job in this location.
  • Calibration comes before anything goes live: the hiring manager rates real applications and sees how the criteria rank the same people, then adjusts the weights.
  • Criteria are edited in Studio and held as versioned settings, so each change is approved and published deliberately and can be traced.
  • A hiring manager sees only their own role categories and jobs.
  • Governance of the criteria stays with you as the employer.

Related: https://www.inploi.com/agents/hiring-agent

Can we see why a candidate scored what they did?

Yes: every score breaks down criterion by criterion, with a score, the evidence from the CV or answers, and a written reason for each, so a recruiter can check how the overall score was reached before acting on it.

  • The overall score is plain arithmetic over the criterion scores and your weights, and the arithmetic is shown. Move a weight and the ranking updates in front of you, with no re-scoring.
  • If an answer could not be used, the score says it was left out, so a gap is never mistaken for a weak answer.
  • The full application and CV stay one click away for a spot check.
  • The reasoning gives your team what it needs to explain a shortlist to a hiring manager or answer a candidate’s question.

Related: https://www.inploi.com/agents/hiring-agent · https://www.inploi.com/security/ai

How consistent is the score if the same application is scored twice?

Scoring the same application again against the same criteria within 90 days returns the same score, because the Hiring Agent stores its result for that long and reuses it rather than asking the model again.

  • A change to the criteria, the weights or the application produces a fresh score.
  • The overall score is fixed arithmetic over the criterion scores, not a second judgment by the model.
  • Before criteria go live, they are trialled on sample applications so recruiters can check that the scores match how they would rate the same people.

Related: https://www.inploi.com/agents/hiring-agent

Does the score compare candidates with each other or against the role?

Each application is scored against the role’s criteria, so a score means the same thing whether a job has eight applicants or eight hundred.

  • Where a job has many applicants, Studio also shows where each applicant stands in that job’s own pool so far.
  • Each job is compared only with its own applicants, never with a different role or location.

Related: https://www.inploi.com/agents/hiring-agent

How do you score roles that don’t ask for experience or a CV?

For entry-level and hourly roles, the criteria are the traits and situations the job involves, such as availability, customer service or working under pressure, assessed from the candidate’s answers rather than from CV history.

  • Screening questions are set per role category and scored by fixed rules; other answers in the application feed the criteria directly.
  • Each role category can weight its criteria differently.
  • Top matches can be invited to a short AI first interview by voice, scored against its own criteria; the transcript is scored, never the voice.

Related: https://www.inploi.com/agents/hiring-agent · https://www.inploi.com/use-cases/quality-in-volume

Could we lose great people with lots of potential to an AI score?

Agents rank and recommend; they never reject, and a person on your team makes every hiring decision.

  • Qualities such as adaptability or service can be written into the criteria and weighted.
  • Candidates may be pointed to roles that suit them better, but anyone can always apply anyway.

Related: https://www.inploi.com/security/ai

Can the AI score appear in our ATS?

Yes, where the ATS takes it: scores always appear in Studio, and the score and its breakdown are written to the candidate in the ATS as a field, so recruiters work in one place. Tribepad, Workday and SmartRecruiters all accept it.

  • For other ATSs, what can be written back depends on what the ATS’s API allows, and it is confirmed during scoping.
  • Where the ATS cannot take the score, recruiters sort and read scores in Studio.

Related: https://www.inploi.com/agents/hiring-agent · https://www.inploi.com/data/integrations

Our ATS already has AI screening. Why add inploi’s scoring?

You may not need to: inploi’s scoring adds most where your ATS’s AI cannot show its reasoning, is not independently audited, or does not see what candidates told the careers site and application before they applied.

  • inploi’s careers site, application and Candidate Agent are useful with or without the Hiring Agent.
  • Where the ATS takes it, including Tribepad, Workday and SmartRecruiters, the score is written to the candidate in the ATS.

Related: https://www.inploi.com/agents/hiring-agent · https://www.inploi.com/use-cases/more-from-your-ats

What if candidates don’t want to deal with AI?

They can say no: every candidate is asked whether they are happy for AI to be used, a no means the Hiring Agent does not score that application, and saying no neither helps nor harms it.

  • An application that is not scored stays in the queue and is reviewed by a person in the usual way.
  • A candidate can also object through the employer at any time, and inploi, acting on the employer’s instruction, stops scoring them.
  • A form version of the application and a route to a person are available.

Related: https://www.inploi.com/security/ai

Agents do the admin. Humans do the hiring.

Let’s find the admin your team shouldn’t be doing.

The hiring experience audit takes four minutes.