As of September 2026, LinkedIn says its AI-powered job search is available to all members worldwide and that classic job search has begun to be gradually retired. The main change is straightforward: instead of relying only on an exact job title and a chain of filters, members can describe the kind of work they want in natural language. The system tries to interpret that intent and match it against millions of job descriptions. Filters have not disappeared; the best way to combine text and filters has changed.
What changes with AI-powered job search
In its official help page for AI-powered job search, LinkedIn recommends being specific about location, experience level, specialty, skills, and industry. The company also explains in its technical overview of the new search that the goal is to move beyond literal keyword matching and understand semantic relationships between what a member asks for and the content of job postings.
In practice, this favors queries that read like a short brief. Instead of searching only for “data analyst”, for example, add where you want to work, the level you are targeting, the skills that matter most, and how recent the postings should be.
The filters that are still available
According to LinkedIn’s current documentation, AI-powered job search results still offer filters for:
- date posted;
- company;
- experience level;
- employment type;
- remote work;
- LinkedIn Apply;
- jobs in your network;
- jobs with fewer than 10 applicants.
These criteria can also be included directly in a natural-language request. The benefit of applying a visible filter after the first search is that it makes a truly important restriction explicit. If you only want jobs posted in the past week, for example, saying so in the query helps the AI understand your intent, while also using the date filter reduces ambiguity.
LinkedIn says other controls are still being brought back, including verified jobs, sorting by most recent or most relevant, and dedicated hybrid or on-site options. The interface may therefore change, and some controls that existed in classic search may not be available in the same form during the transition.
How to write a useful request without turning it into a long paragraph
A good query usually contains four to six components. You do not need all of them every time, but the more important a criterion is, the earlier it should appear.
- Role or function: state the job you want and, when useful, include closely related titles. Example: “data analyst or BI analyst.”
- Location: specify a city, state, country, or geographic range. For remote work, make clear whether you are open to employers anywhere or only roles hiring in your country.
- Experience level: use terms such as internship, entry level, mid-level, senior, or an equivalent description of your experience.
- Skills or specialty: mention two or three capabilities that genuinely distinguish the search, such as SQL, Power BI, B2B sales, or enterprise account management.
- Industry or context: add the type of company or sector when it meaningfully changes the results, such as healthcare, fintech, retail, or SaaS.
- Practical constraints: include recency, employment type, remote preference, or criteria such as “in my network” and “under 10 applicants” when relevant.
Query templates to adapt
The examples below work best as starting points rather than rigid formulas. Replace each element according to your priorities.
- Role + location + skills: “Entry-level data analyst jobs in Recife or remote in Brazil, requiring SQL and Power BI, posted in the past week.”
- Function + industry + experience: “Customer success jobs at technology companies in Brazil, mid-level, with B2B account experience and English.”
- Two related functions: “Business development or partnerships opportunities in fashion, with intermediate experience and remote work options.”
- Network + competition: “Senior finance manager jobs in my network with fewer than 10 applicants.”
- Career transition: “Operations or project management jobs that can use experience in customer service and process organization.”
When a request includes a criterion that does not yet have a dedicated filter, such as a very specific hybrid-work preference, treat the wording as a signal for the AI and confirm the condition in each job description.
When to use text and when to use a filter
| Criterion | Best use | How to review it |
|---|---|---|
| Role, specialty, and skills | Describe them in the query | Check whether the results represent roles that are genuinely close to your goal |
| Location | Describe it clearly in the query | Confirm city, hiring eligibility, and remote rules in the posting |
| Experience level | Query + available filter | Read the requirements because level labels vary across employers |
| Date posted | Query + available filter | Use the filter when recency is critical |
| Employment type and remote | Query + available filter | Confirm contract type and work model in the posting |
| Company, network, and under 10 applicants | Use filters when they are genuine constraints | Do not treat applicant count as a measure of job quality |
Why you should still review the results even after writing a strong query
LinkedIn itself acknowledges that search results do not always match what members are looking for and that it is continuing to improve relevance and restore filters. The platform also says it may use profile data and search activity to personalize results. Two people can therefore use similar requests and see jobs in a different order.
AI also does not replace reading the job posting. Before applying, confirm at least the actual role, location, work arrangement, employment type, seniority, mandatory requirements, and posting date. When a skill is listed only as preferred, do not automatically treat it as a disqualifying requirement.
A simple workflow for searching without losing precision
- Start with a specific but not excessively narrow request. Combine role, location, level, and two or three important characteristics.
- Look for patterns in the first results. If the wrong functions appear, adjust the role; if the geography is too broad, rewrite the location.
- Apply available filters to hard constraints. Date, experience, employment type, remote, company, network, and applicant count are good candidates for filtering.
- Refine one dimension at a time. This makes it easier to see which change improves or worsens the results.
- Read the posting and give search feedback. LinkedIn provides positive and negative feedback controls on results to help improve the experience.
The key point: natural language does not mean giving up criteria
The new search is most useful when your request expresses the intent that used to be split across keywords and filters. Current filters still matter for objective conditions. The safer combination is to use natural language to explain the kind of opportunity you want and filters to reinforce the constraints that cannot be missed. Because classic job search is being gradually retired and some filters are still returning, check LinkedIn’s official documentation whenever an important control disappears or changes name.
