To get started with AI-Assisted Search and Projects in Recruiter, click the search bar and select Start a new AI search or Create a new AI project.

You can use AI-Assisted Search and Projects to start a search, create a project, or ask a question related to your Recruiter contract or experience. With AI-Assisted Search and Projects, you can let LinkedIn know your hiring needs in your own words. LinkedIn Recruiter will automatically create and optimize a candidate search for you with smarter suggestions on search filters, allowing you to interact with search in a much more intuitive way. With AI-Assisted Search and Projects, you can accelerate your hiring efficiency and uncover new hiring strategies by expanding or narrowing your search results through smarter suggestions on search filters.

Here are some examples of what you can do with AI-Assisted Search and Projects: 

• You can create a project by typing “Help me create a project for hiring Software Engineers in the Bay Area.”
• You can jump start your sourcing by entering details such as your preferred job titles, locations, years of experience, skills, companies, industries, etc. For example, you can enter “Find a Senior Product Manager in New York with 6 years of experience.”
• Add a job description or job link in the search bar to find candidates that meet the role’s requirements.
  • This feature is available only to LinkedIn Recruiter and LinkedIn Recruiter Professional Services Plus (RPS+) customers. It isn’t available to LinkedIn Recruiter Professional Services customers.
• You can also ask questions to get recommendations to refine or expand your search, such as “What are some companies that I can add to my search?” or “What are automated pipeline stages?”

AI-Assisted Search will provide you with a list of candidates that match your requirements.

For LinkedIn Recruiter and LinkedIn Recruiter Professional Services Plus (RPS+) customers with access to Advanced AI Search, AI-Assisted Search will display an AI summary on candidates’ profile cards detailing how they’re a fit.  Each candidate’s profile card includes a relevance label that indicates how well their qualifications match your criteria:

• High qualification relevance indicates 70% or more qualifications matched
• Medium qualification relevance indicates 30% or more matched
• Low qualification relevance indicates less than 30% matched.

Click the label to see examples of how the candidate aligns with your qualifications.

To learn more about AI-Assisted Search and Projects, see Using AI-Assisted Search and Projects in LinkedIn Recruiter.

Using this feature is optional. You are still able to initiate a search or create a project without using AI-Assisted Search and Projects. To start a search without using AI-Assisted Search, click the search bar on your Recruiter homepage and select Start a Recruiter search with filters from the dropdown.

You also have the option to set the traditional keyword search as default under AI Preferences in Product settings. If you default to the traditional keyword search, all searches will default to keyword and AI will be available as a secondary option. 

AI-Assisted Search tries to understand your hiring intent (e.g., find a Public Relations Associate in Texas) and breaks it down into the most relevant search filters (e.g, job title and location) for you to find the right people.

Note: AI-Assisted Search allows you to express your hiring intent in a more intuitive way, while the underlying search algorithm remains the same.

You can learn about AI-Assisted Search and Projects by accessing the AI-Assisted Search and Projects in Recruiter Help Center article. You can also access training on LinkedIn’s AI features through LinkedIn’s Learning Center.

To access training on LinkedIn’s AI features in Recruiter:

Here is a full list of search filters that you can use:

• Job title
• Location
• Open to work 
• Active talent (candidates who are more likely to respond (looking for a job))
• Internal candidate
• Interested in company 
• Company connections (candidates who know people at my company)
• Skills and assessments
• Companies currently working for
• Schools
• Year of graduation
• Industries
• Seniority 
• Who I’ve messaged
• Who I haven’t messaged
• Candidates that aren’t in any projects
• Spoken languages
• Network relationship (candidates I’m connected to or have common connections with)
• Years of experience 
• Years in current company
• Years in current position
• Fields of study
• Degree
• Company size (works at a company of a specific size)
• Past applicants
  

Here is a list of search filters that you can’t use:

• Race
• Gender
• Nationality
• Sexuality
• Age
  

Yes, you can still do Boolean searches in Recruiter. If you want to perform a Boolean search without any AI assistance, add your Boolean string to the search bar and you’ll be given the option to run a traditional keyword search from the dropdown. You can also use AI-Assisted Search to create and modify Boolean strings. If you enter a Boolean string to the search bar, AI will search with the Boolean string. Alternatively, you can ask AI to create a Boolean string for a specific open role with conversational language. For example, you can enter “Write a Boolean that consists of recommended skills for Java Developers.”

Advanced AI-Search goes beyond matching on keywords and can now match on more hard-to-define, critical skills such as “data-driven decision maker”. With Advanced AI Search, add, edit, or delete qualifications in the filter pane on the left, and view an AI summary on the candidate profile card detailing why they’re a good fit. Each candidate’s profile card includes a relevance label that indicates how well their qualifications match your criteria: 

• High qualification relevance indicates 70% or more qualifications matched, 
• Medium qualification relevance indicates 30% or more matched, and 
• Low qualification relevance indicates less than 30% matched.
  

Note: Advanced AI Search is currently available to LinkedIn Recruiter and LinkedIn Recruiter Professional Services Plus (RPS+) customers whose language setting is English. It isn’t available to LinkedIn Recruiter Professional Services customers.

AI-Assisted Search in LinkedIn Recruiter is a generative AI (GAI) feature that will save time by enabling you to quickly search for candidates and get recommendations just by stating what you need in your own words into a prompt box. 

With Advanced AI-Assisted Search, we expand on AI-Assisted Search (standard) capabilities to go beyond simply matching on filters. Advanced AI-Assisted Search matches hard-to-define critical skills not typically listed on candidate profiles or resumes, but are often listed on job descriptions, such as “has experience solving ambiguous problems” or “can lead large, cross-functional workstreams.” By better understanding a recruiter’s intent, Advanced AI-Assisted Search delivers better search results and provides more thorough qualifications and skill representation for candidates.

With Advanced AI-Assisted Search, you can start a search for candidates in even more ways: 

• Paste in intake notes from a hiring manager
• Type in your hiring goals
• Paste a job description or job link directly into the search bar to find candidates who meet the role’s requirements
  

Advanced AI-Assisted Search enables you to use plain text language which then gets translated via AI into required and preferred qualifications, alongside existing filters (e.g. title, industry, schools, companies, and locations) to obtain more accurate search results. After the initial search, you can easily review which qualifications were met and where gaps exist for each candidate.

With Advanced AI-Assisted Search, you can add, edit, or delete qualifications, and view an AI summary on the candidate profile card detailing why they’re a good fit. Each candidate’s profile card includes a relevance label that indicates how well their qualifications match your criteria: 

• High qualification relevance indicates 70% or more qualifications matched, 
• Medium qualification relevance indicates 30% or more matched, and 
• Low qualification relevance indicates less than 30% matched.

Note: Advanced AI-Assisted Search is currently available to LinkedIn Recruiter and LinkedIn Recruiter Professional Services Plus (RPS+) customers whose language setting is English. It is not available to LinkedIn Recruiter Professional Services (RPS) customers.

At this moment, there is no limit to the total number of questions you can ask. However, within 1 minute, you can ask no more than 10 questions.

We standardize profile titles, skills, and schools across many languages, so if you’re searching for titles, skills, and other facets, we will return matching profiles regardless of the profile language. For resume search, we currently only process and standardize resumes written in English.

There are two ways to see the previous search results in AI-Assisted Search: 

• Click the  More icon at the top of the tool and select View search history.
• Click Show filters at the top of the tool and manually change the search filters to reflect what you would like to see in search results.
  

No, we don’t currently support navigation actions. However, you can message a candidate or add them to your pipeline from the candidate’s profile card.

No, you need to make any updates to a project manually from within the project itself. 

If you believe AI-Assisted Search and Projects misunderstood or misinterpreted your question, you can rephrase or change your question. Additionally, you can click Show filters at the top of the tool to expand the filter view and manually add or change your filters.

AI-Assisted Search and Projects will provide you with suggested prompts you can select to help guide your search. You can provide in-product feedback by clicking the thumbs up or thumbs down icon if the results do not meet expectations so the product can learn from this and improve over time.

If you want additional changes or refinements to your original request, you can make an advanced search at any time. Click Show filters at the top of the tool to expand the filter view, and click Advanced search to see additional filters.

To clear your search click the  More icon at the top of the tool and select Clear search. This will clear all search filters so you can get the most relevant suggestions and results when you enter a new search in the AI-assisted search tool.

If you want to make additional changes or refinements to your original request, click Show filters at the top of the tool to expand the filter view, and click Advanced search to see additional filters.

However, if you want to start a completely new search, or change all the criteria, we recommend that you clear your search or start a new search from the global search bar so that you get the most relevant suggestions and results. To clear your search click the  More icon at the top of the tool and select Clear search.

When possible, AI-Assisted Search uses facets to construct your search since facets give the most precise results. For skills, a faceted search will match many terms that match to a given skill (e.g. the terms “software product development”, “software development”, “software engineering”, and “sw development” will all map to the Software Development skill). A faceted search also attempts to return only users who possess that skill whereas a keyword search returns users who have that term anywhere on their profile which can lead to irrelevant results.

We don’t currently offer a way to save and re-use search requests. Because we are constantly evolving the capabilities of our systems, you may not always see the same search results from a given request. 

Once you’ve created a search, you can save that search to a project to keep working within the same search across login sessions or to share that search with other users on the contract. To save your search, click the  More icon at the top of the tool and select Save as new search.

Inputting a request in AI-Assisted Search helps streamline your searching so you don’t have to spend time navigating and manually updating filters. Simply enter your request into the AI-assisted search tool to refine your search. 

AI-Assisted Search and Projects will provide you with suggested prompts you can select from to help guide your search. You can also ask it for suggestions by entering “Help me expand my search.”

AI-Assisted Search translates natural language into a search for you using a faceted search and keyword search. 

Faceted search (i.e. filters) standardizes members' profiles and returns candidate matches. Faceted search will match many terms that match to a given skill (e.g. the terms “software product development”, “software development”, “software engineering”, and “sw development” will all map to the Software Development skill).

Keyword search returns members who have that term anywhere on their profile. To help control keyword search results, you can use Boolean operators. 

If AI-Assisted Search populates the faceted and keyword search with the same terms you would do manually with traditional search, the results will not differ. The only time you may see different results is if AI-Assisted Search translates natural language to different faceted search and keyword search terms.

You can scroll up within your search session to see past chats from your current session. Once the session ends, the history is no longer visible. If you are working within a project, your chat history within that project will be preserved across login sessions.

You can see how users on your Recruiter contract are using AI-Assisted Search in the Usage report. 

In the Usage chart, the Avg. daily active users and Avg. searches performed tabs show data for overall search, AI-Assisted Search, and non-AI search. 

In the Usage Details table, the AI-Searches column displays the number of new AI searches performed during the selected timeframe. The AI-Searches % column displays new AI searches as a percent of the total new searches performed during the selected timeframe.

You can ask for the following types of recommendations in AI-Assisted Search and Projects, depending on your license type.

If you have LinkedIn Recruiter and LinkedIn Talent Insights, you can dig further into surfaced insights by clicking on the View Talent Insights button at the top right of the search within an existing project. You can also enter “Show me a report for this Talent Pool” or “Generate a report for this talent for major cities on the West Coast” into the chat box to generate a Talent Insights report.

AI-Assisted Search and Projects enables recruiters searching for potential job candidates to provide a plain language input into the search tool in LinkedIn Recruiter, which then gets translated via AI into a structured search to leverage Recruiter’s existing search functionality and filters (e.g., title, industry, schools, companies, and location) to obtain search results. For example, a prompt like “Marketing Managers in Dallas with sales enablement experience” would create the following search filters: job title (Marketing Manager), location (Dallas, Texas), and skills (sales enablement).

Recruiters can manually adjust the search filters or try different inputs to reach the desired outcomes. They can provide in-product feedback by clicking the thumbs up or thumbs down icon if the results do not meet expectations so the product can learn from this and improve over time.

Our models that understand intent and translate that intent into a search are continuously evolving. LinkedIn will continue to improve the types of requests we support and improve our ability to deliver on recruiters’ desired search outcomes.

The generative AI (GAI) models currently used to power AI-Assisted Search and Projects in Recruiter were developed by OpenAI and LinkedIn uses these models through Microsoft’s Azure OpenAI Service. If you wish to learn more about how these models were trained, please see OpenAI and Azure OpenAI Service.

AI-Assisted Search and Projects in Recruiter have been subject to our review process to ensure it can be used in compliance with applicable laws, our privacy policy, and our contracts with customers. Learn more about how LinkedIn is operationalizing its Responsible AI Principles. Our security commitments with respect to customer data are set forth in our agreement with customers. Like all of our services, AI products are subject to security reviews designed to prevent harm.

The generative AI (GAI) models currently used to power AI-Assisted Search and Projects in Recruiter were developed by OpenAI, and LinkedIn uses these models through Microsoft’s Azure OpenAI Service. If you wish to learn more about how these models were trained, please see OpenAI and Azure OpenAI Service.

Customers' inputs into this GAI feature are not sent to OpenAI, and Microsoft has committed not to use them for further training of AI models.

Recruiter data is used to improve AI-Assisted Search and Projects in two ways:

• We look at searches, particularly ones with feedback, to help us understand where we can improve our translation of input into a search and to identify areas to prioritize for product improvements. This is similar to what we've always done with search data.
• As with other search data, we constantly run experiments to improve search ranking and retrieval and to understand how changes to our models improve outcomes for search. For instance, we might make changes to the way we create queries from user input and we will run an experiment to see if this is leading to more successful conversations between recruiters and candidates.

We may also use this data in the future to train in-house models, for example to classify whether a request is a search creation request or a project creation request.

LinkedIn's economic graph (skills, companies, schools, etc.) is constantly being updated to include new entities. Members decide what information to share and update on LinkedIn, and LinkedIn Recruiter uses member profile and LinkedIn graph information to retrieve search results.

For AI-Assisted Search and Projects, we regularly expand our evaluation data set to include new examples and use this data to evaluate different model versions to ensure optimal performance.

The LinkedIn Recruiter product provides a ranked list of candidates corresponding to a search query. Given a search request, candidates that match the requested keywords and facets are selected and then ranked based on various factors. This includes signals such as the similarity of their work experience/skills with the search criteria, and the likelihood of a response from an interested candidate, which are weighted using machine-learning models.

As a recruiter, you choose what qualifications are relevant for open roles. LinkedIn Recruiter does not play a role in decisions made on job applicants by a recruiter, but instead plays a role in a recruiter's efforts to increase applicant pools by reaching out to LinkedIn members to encourage them to apply for specific roles. Members choose how much information they wish to share with LinkedIn. Members can highlight certain experiences and skills, making them more likely to be responsive to recruiter search efforts.

No. AI features in Recruiter are primarily recommendations and do not make hiring or sourcing decisions for the users. Users will still own the decision making of engaging or not engaging candidates on LinkedIn Recruiter. For example, AI-Assisted Search and Projects enables recruiters searching for potential job candidates to provide a plain language input into the search tool in LinkedIn Recruiter, which then gets translated via AI into a structured search to leverage Recruiter’s existing search functionality and filters (e.g., title, industry, schools, companies, and location) to obtain search results. For example, a prompt like “Marketing Managers in Dallas with sales enablement experience” would create the following search filters: job title (Marketing Manager), location (Dallas, Texas), and skills (sales enablement).

Once search filters and results appear, recruiters can manually adjust the search filters or try different inputs to reach the desired outcomes. Users also have the ability to provide in-product feedback, via a thumbs up/down, when the results do not meet their expectations so the product can learn from this and improve over time.

Our product does not screen out candidates who have applied to jobs or make hiring decisions. A customer’s own team decides who to reach out to and who to advance in the pipeline once a person has applied.

LinkedIn strives to ensure that its products and services are fair, including by measuring and mitigating algorithmic bias. Our objective is to ensure equally qualified members receive equal treatment by our models.

We don’t enable the filtering of specific gender or ethnicity groups for hiring and this feature does not change that. AI-Assisted Search and Projects enables a recruiter searching for potential job candidates to provide a plain language input into the search tool in LinkedIn Recruiter, which then gets translated via GAI into a structured search to leverage Recruiter’s existing search functionality and filters (e.g., title, industry, schools, companies, and location) to obtain search results. We don’t provide search filters that target specific gender or ethnicity groups for hiring; therefore, this feature does not enable that targeting.

Users cannot opt-out of AI-enabled features, but can choose to use manual inputs instead. For AI-Assisted Search and Projects, a recruiter can choose to use manual keyword or Boolean searches instead of the AI-Assisted Search functionality. To have more control over their experience, they can also turn off the defaulted AI-Assisted Search experience by updating their AI preferences in their Product settings.

This feature is powered by “artificial intelligence” or “AI.” Your use of this feature is subject to our User Agreement and Privacy Policy, and you must continue to abide by our Professional Community Policies. Below we have highlighted some of the key points associated with your use of this feature:

Quality and Safety: While we try to provide an AI feature that you will find useful, it is possible that the output (or answers) it generates can be factually wrong, misguided, and/or not useful. It’s important that you assess the output and answers of this feature for accuracy and appropriateness for your purpose and that you do not solely rely on it.

Feedback and Human Review: We value your feedback. If you choose to provide us feedback, we can use it to help make the feature better over time for all members. Human review of the “session” and/or your inputs into this feature and our related responses is possible when you provide feedback or otherwise. To build, train, secure, and improve the accuracy and quality of this feature and our Services (as defined in our User Agreement) - even when you do not provide feedback - we may manually review your inputs and the responses we provided.

Limits: While we may make this feature available to a subset of members, we reserve the right at any time to limit access and use, change the feature, and/or cease providing the feature, including for reasons such as abuse, misuse, or overuse of the feature.

Privacy: Information that you provide as inputs, the responses, and your use of this feature will be associated with your enterprise seat and will be considered “customer personal data” under our Data Processing Agreement. While your prior “sessions” with this feature may not be viewable by you after the close of a session, we will retain this information beyond the session. Like other personal data on LinkedIn, the purchaser of your enterprise seat can request to see what we have retained and request to take action on your personal data.