When the AI evaluation lands in the ATS, what happens to the assessment workflow?

On 14 October 2026, iCIMS announced via PR Newswire that LinkedIn Recruiter System Connect Plus for Hiring Assistant will become available for the iCIMS applicant tracking system. The integration is expected to launch in spring 2027 for customers who use LinkedIn Hiring Assistant and hold a LinkedIn Recruiter licence. According to the release, applicant data from iCIMS is synced into LinkedIn Hiring Projects linked to a specific requisition in the ATS. Hiring Assistant then evaluates applicants against the hiring criteria by combining insights from the LinkedIn profile with resume data held in iCIMS. The evaluations sync back to iCIMS in real time, so recruiters can surface applicants based on the evaluation result without switching systems.

LinkedIn estimates that recruiters will save an average of 1.5 hours per role on shortlisting. That is a vendor estimate rather than a measured outcome, and the release itself presents it as an expectation.

On the surface this is a CV screening story, not an assessment story. But it illustrates a shift that anyone responsible for assessment should watch closely. When an AI evaluation becomes a field in the ATS that can be sorted and selected on, it moves the point in the process where decisions are actually made. The test score has traditionally been the first structured data point about a candidate. Now another structured data point can arrive earlier, based on profile and CV, and in practice it will decide who receives a test invitation at all.

It is worth noting that the release does not mention assessments, tests, skills, bias or compliance. The closest it comes is the phrase about evaluating applicants against hiring criteria. That is not a criticism of iCIMS or LinkedIn. It is simply a statement of what each organisation will need to work out for itself before switching the integration on.

What is the evaluation based on?

According to iCIMS, the evaluation combines LinkedIn profile insights with resume data in the ATS. Both are self-reported. A test, by contrast, measures behaviour or ability under controlled conditions. If the AI evaluation is used to decide who progresses to testing, you need to know which criteria it applies and whether those are the same criteria the test has been validated against. If they differ, the candidates the test was designed to identify may never reach it.

Where in the funnel does it sit?

Once the evaluation syncs back in real time and applicants can be surfaced by it, it easily becomes a de facto screening stage. That can mean candidates are filtered out before the test has contributed anything. With large applicant volumes that may be a sensible prioritisation, but it should be a deliberate decision, documented in the workflow, rather than a side effect of an integration that was adopted to save time.

How are two evaluations weighed?

An ATS holding both an AI evaluation and a test score for the same candidate raises the question of what happens when the two point in different directions. Without a rule for which carries more weight, the decision falls back on the individual recruiter's judgement. That is precisely the inconsistency structured tools are meant to reduce. The more automated evaluations land on the same candidate record, the more important it becomes that sequence and weighting are settled in advance.

The integration is not due until spring 2027, so there is time to prepare. Start by mapping which fields in your ATS are currently used to reject and which are used to rank. Decide whether the AI evaluation should sit before or after the test, and write down the reasoning. Make sure test results continue to sync into the ATS in a structure where they are as visible as the new evaluation, so they are not buried beneath it. And remember that AI systems used to select and evaluate candidates fall under the EU AI Act as high-risk. Documenting how an evaluation feeds into a decision is therefore not a formality but part of the requirements.

The iCIMS and LinkedIn announcement is, in itself, an integration story. What it means for the assessment workflow depends entirely on where each organisation chooses to place it.

Kilde: https://www.prnewswire.com/news-releases/icims-integrates-with-linkedin-hiring-assistant-to-deliver-a-more-connected-experience-and-help-recruiting-teams-hire-quality-talent-302907896.html