The AI copilot for recruiters

A recruiter's week fragments into work that requires judgement and work that only requires time. Reading a hundred CVs to discard eighty is time. Deciding which of the remaining twenty to call is judgement. Automating the first without touching the second is where the hours come back.

In practice the highest-yield candidates for automation are the ones that repeat identically: extracting structured data from CVs in different formats, first-pass screening against explicit criteria, scheduling with several calendars, and writing the interview summary from notes. None of those decide anything; all of them consume afternoons.

The temptation once screening works is to extend automation into the decision, and that is where recruiting processes lose their credibility. The reject is the highest-stakes output of the process for the person receiving it, and it should never be produced by a system without a person behind it.

There is a second reason beyond fairness. A recruiter who reads the borderline cases learns what the hiring manager actually wants, which is rarely what the job description says. Removing that reading loop makes the process faster and the requirements permanently wrong.

The tempting metric is time saved, and it is the least useful, because it measures the tool rather than the outcome. A screening that is twice as fast and produces worse shortlists is a loss dressed up as a gain.

The pair worth watching is shortlist precision and time to first qualified interview. Precision is the share of presented candidates the hiring manager accepts as interview-worthy; if that rises while the calendar time falls, the assistant is working. If precision falls, the criteria need rewriting before anything else is tuned.

Open the LinkedIn profile you want to keep. Open the Hice extension and sign in, then choose to save a candidate. Review the preview: name, role, experience and other available information must belong to the right person. Complete required fields and confirm the save.

The operation concerns one profile at a time and starts with your choice. Browsing LinkedIn does not automatically collect lists. Once saved, the record is in your Hice database: add notes, a CV or a search link without retyping the profile from scratch.

Separate necessary requirements from preferences: skills, experience, location and verifiable constraints. Ask the assistant to search accessible Hice data and show evidence and missing information for each candidate. Availability should not be inferred merely because a profile is public.

Connector search and model analysis are not automatically the same engine as Hice's native matching. Check available operations before requesting a function. If an attachment or semantic search is inaccessible through the connector, use the app's supported flow rather than expecting the model to read content it has not received.