Define the role before automating the pipeline
Write down the outcomes expected in the first three months, the skills required to achieve them and what can be learned on the job. Separate genuine requirements from preferences. A long list of optional technologies can discourage qualified applicants and produce a weak screening rubric.
For a support role, for example, the outcomes might be resolving a customer issue clearly, recognising when to escalate and documenting the next step. Each outcome needs evidence: a scenario response, a relevant example or a short work sample. Decide how reviewers will assess that evidence before applications arrive.
Publish a complete job and explain the process
Give applicants a clear title, full responsibilities, employment type, work arrangement and workplace or eligible countries. State compensation when it is confirmed. “Remote” describes where work happens; it does not tell a candidate whether the employer can hire in their country.
Explain the application stages and whether an AI interview is involved. Tell candidates what they need, how their information is used and where to ask for an alternative format. In Recruit, the public careers page and individual job pages connect the job description to the application form.
Use resume screening to organise evidence
Compare applications against the criteria for this role, rather than a generic ideal-candidate profile. Treat missing information as missing information. A resume that does not mention a skill is different from evidence that a candidate cannot perform it.
Recruit can score resumes against a role and, when the employer enables it, invite candidates who meet the configured threshold. Review borderline cases and check your automation settings. A threshold is a workflow setting, not proof that the underlying assessment is correct.
Run structured AI screening interviews
Give candidates comparable opportunities to demonstrate the same competencies. Use a common core of questions, with follow-ups that clarify their answers. Ask about choices and trade-offs rather than trivia that has little relationship to the work.
An AI interviewer can collect responses on a candidate’s schedule and produce transcripts and scorecards. In Aiinak Recruit, Iris supports screening and technical interviews within the ATS workflow. The hiring team should inspect the underlying responses whenever a score or recommendation influences the next step.
Review the shortlist and improve the process
Record the reason for advancing each candidate, the evidence supporting it and any uncertainty to explore in the next interview. Do not let a concise generated summary replace a relevant answer in the transcript. If automated pipeline decisions are enabled, assign someone to review the configuration and handle requests for human review.
Track application completion, interview completion, time to review and how often reviewers change an initial recommendation. Look for recurring technical failures or questions that confuse applicants. Improve the questions and criteria before increasing automation. These checks reveal whether the process helps your team rather than simply moving candidates faster.
For candidate information, read the AI screening notice and privacy policy.