LinkedIn Is Retiring Classic Job Search. Your Profile Is Now Part of the Query.

LinkedIn is retiring classic job search. Its own help page says so plainly: AI-powered job search is available to all members worldwide, and the company is "gradually retiring classic job search starting in September." If you've opened the Jobs tab lately and found a chat-style box where the keyword field used to be, that's why.

Most of the commentary so far is about whether the new search is better or worse. That's a fair question, but it isn't the useful one if you're actually looking for work. The useful question is what you should change now that the search reads meaning instead of matching words. Because the answer touches your search habits and your profile.

What actually changed

Here's what LinkedIn's help article says, in short:

• You describe the role in your own words. Its examples include "Entry-level sales jobs in healthcare posted in the last week" and "Finance manager roles in my network with less than 10 applicants."
• The search is powered by large language models that LinkedIn says are fine-tuned on two decades of its data. It's meant to understand intent, so "entry-level marketing jobs" can bring back a "Marketing Coordinator" listing that never uses the phrase entry-level.
• Some filters stay: date posted, company, experience level, employment type, remote, Easy Apply, in my network and under 10 applicants.
• Others are missing for now. LinkedIn says it's working to bring back Verified job, the Most recent / Most relevant sort, and Hybrid / On-site.
• Suggested filters appear under the search bar and adapt to what you typed.
• LinkedIn "may use data from your profile and search activity to personalize your results."

That last line is the one I'd underline. The search isn't only reading your query anymore. It's reading you.

Your profile is now part of the query

Under keyword search, your profile mattered mostly to recruiters searching for you. Now LinkedIn says it may use your profile to shape the jobs you see. So a stale headline doesn't just hide you from recruiters. It can also tilt your own results toward the job you had three years ago.

That changes what a profile refresh is for. Before you run a single search, check three things:

1. Headline. Does it describe the work you want next, or only the title you hold? If you're moving from sales operations into revenue analytics, a headline that says "Sales Ops Manager" and nothing else is pointing the system the wrong way.
2. About section. Write two or three plain sentences about the problems you solve and the kind of team you want to join. Plain language helps here. A model reading for meaning gets more from "I build the reporting sales leaders use to forecast" than from a stack of buzzwords.
3. Skills and recent roles. Make sure the skills that matter for the next job are actually listed, and that your current role's description mentions them in context.

None of this is new advice. What's new is that skipping it now quietly hurts your own search, not just your visibility.

Write searches like a brief, not a wishlist

LinkedIn's own guidance is to "be specific," including location, experience level, specialty, skills or industry. People testing it say the opposite failure is real too. Career writer Kevin D. Turner, as quoted in a post by Christine Westbrook on LinkedIn, reported that the more detail he added, the worse his results got, and that trimming the query helped. That's one person's testing, not a study, but it matches how these systems usually behave: a query with eight conditions gives the model eight ways to misread you.

So here's the approach I'd take:

• Start with three things: the role, the place, and one thing that really matters (industry, seniority or company size).
• Use the remaining filters for the hard rules. If you must have remote work or fewer than ten applicants, those filters still exist. Let the filter enforce it instead of hoping the model picks it up from a sentence.
• Change one thing at a time. If results drift, remove the last detail you added before you add another.
• Save the searches that work. You'll want them as alerts once the results look right.

A hypothetical example: say you're a product designer in Bengaluru who wants fintech work. "Senior product designer roles in fintech in Bengaluru" is a better first search than a paragraph listing your tools, your preferred team size and your salary band. Get a decent list, then use filters for remote and date posted.

Watch for what's missing

The missing filters matter more than they look. Without a Most recent sort, it's harder to spot new postings the day they go up, and applying early has always been one of the few edges a job seeker controls. Until that sort comes back, two habits help:

• Put timing in the query ("posted in the last week" is one of LinkedIn's own examples), or use the Date posted filter, which is still there.
• Keep the Under 10 applicants filter in your rotation for roles you really want.

The Verified job filter being gone for now is worth noting too. Until it returns, give a listing a quick sanity check before you send personal details. Look at the company page, the poster's profile, and whether the role also appears on the company's own careers site.

Recruiters are changing their tools too

The job seeker side isn't the only thing moving. LinkedIn announced Hiring Assistant 2 for recruiters in late September, and ERE's coverage of its Talent Connect keynote says the agent fills in sparse searches using a recruiter's past hiring history, and adds "fit signals" and a "candidate interest" estimate for each recommended person. According to ERE, that interest estimate draws on platform engagement, Open to Work details and how often a person answers InMail.

So on both sides, the system is reading context, not just keywords. That's a good argument for keeping your Open to Work preferences accurate and for actually replying to recruiter messages, even with a short "not right now." Silence may count against you in ways you can't see.

What I'd do this week

If you're job hunting, or might be in the next six months:

1. Rewrite your headline and About section for the job you want next.
2. Run three short searches and save the ones that return sensible results.
3. Use filters for the rules you won't bend on.
4. Keep Open to Work preferences current, and answer recruiter InMails.
5. Check the help page now and then. LinkedIn says it will keep updating it as filters return.

I don't think AI search is automatically better or worse than the old version. It's different, and it rewards people who treat their profile as part of the search. That's a small change in habit with a real payoff, and most people won't bother making it.

Sources

• LinkedIn Help: Discover new opportunities with AI-powered job search
• ERE: LinkedIn's Hiring Assistant 2 Aims to "Just Be Smarter"
• Christine Westbrook on LinkedIn, sharing Kevin D. Turner's testing: LinkedIn's AI Job Search: More Detail, Worse Results?