Google job search function: AI has much to learn


In June, Google released its new AI-powered job search function. It searches major job posting sources (such as Indeed, LinkedIn, Careerbuilder, Glassdoor, and similar sites), as well as individual employer and staffing agency websites. With Google’s powerful tracking of data, what’s not to love about finding so much information using one web tool? Right?

It turns out that the Artificial Intelligence (and presumably the Google developer pool) has much to learn to make this a reliable and time-saving tool for jobseekers.

Mystery of late “applies”

In the weeks since the AI function was released, we have received resumes from individuals applying for jobs no longer available. These hadn’t even been advertised on our website for a month or longer. We’ve been documenting these occurrences, hoping to recognize a pattern. Today, we’ve solved the mystery. The pattern is now clear: each person identified Google Search as their source for finding the job advertisement.

Today I went to Google myself. I entered “Frank’s Employment Jobs” as the search criteria and examined the results. It displayed three positions prominently, but indicated that we actually have 64 more openings advertised (see the green arrow in the photo).

Uh oh. What the search results displayed when I clicked on the 64 included all of the jobs we have advertised during the period since the AI release. Even those unpublished from our website when the searches were completed – more than a month ago.

I clicked on a couple of these job advertisements from openings earlier in the summer. Then clicked again to “View On Job Portal.” Google had correctly cached our original posting on our website, including description of the opportunity, the salary range, the location, and more. The only way to know that this is no longer a live search is if someone happens to notice the “Posted Date” field is blank.

Google’s AI robustly searches, finds, and stores employment advertising, gleaning information from a nice variety of sources. This has the potential to offer even more powerful search consolidation tools than Indeed! However, the AI hasn’t yet learned that job openings are timing-specific and should be cleared from its memory when no longer published. That’s pretty frustrating for you as a job seeker, applying for what seems a current job opening, only to find that the position has long-since been filled.

Adjusting while Google AI learns

While we’re waiting for the AI to learn about the timing-specific aspects of advertised openings, we recommend that human jobseekers adjust expectations for search results on Google. Be aware that Google’s cache is storing lots of data, but not (yet) deleting it. We recommend that you double-check the job board, employer site, or employment agency website to be sure that the search is still active.

For our active direct-hire and temp-to-hire openings, as well as some of our temporary staffing opportunities, feel free to bookmark our Jobs Portal. This is synchronized with our database. When the hiring process is completed or the resume pool is closed, our job listings are immediately refreshed.

2 Responses to “Google job search function: AI has much to learn”

  1. Elyse

    A big shout out to Jason W, who responded to my message yesterday about these Google issues. He suggested some tweaks that we can make in coding ~ we’re working on that! So it’s not just AI that’s learning to better communicate digitally. I’m learning too! =)

  2. Elyse

    A big shout out to Hannah R, who stopped by today after applying for one of those older inactive posts through Yahoo on her phone. The website sent her a confirmation, but didn’t forward her email to us. So-o-o-o, we’re working on this mystery too. Hope to have resolutions soon.


Leave a Reply

  • (will not be published)