One of my biggest challenges is trying to find the best way to use the AI assisted search feature using a job description. I notice I am not able to find a close match to the type of talent I’m looking for. 

Hi ​@wmann, what a great question! 

I do know our engineers were/are working on a way to enter a Job Description into the AI-Assisted search box on the home page of Recruiter. What I don’t know is whether that has ‘ramped’ yet, and, if so, when it is going to be fully launched to all our customers. We roll out updates incrementally to be able to track how they are working, etc. 

It’s a pretty cool feature. There will be a character limit, so if they Job Post is very long, you will need to cut it down.

What I recommend is: get the most important details from the Job Post (skills, location, experience, etc.) and use that as your starting point. (“Please find me a ___ with X skills and Y years of expereince in location.”) Then use that in the search field to start your search.

If you would like to come to a Live Q&A, we can work on this together. Here is the registration link for all the sessions:https://training.talent.linkedin.com/office-hours-general-qa/653597. Mine are on Wednesdays at 1:00pm Pacific. I’d love to brainstorm this with you!

​@Nicole Rodney-LinkedIn  thank you so much this is helpful, and I will definitely sign up for the Live Q&A training, they have been super helpful. It’s great to get live in real time help. I also notice that when I use the JD with just the bare bones it will bring up unrelated roles and I often need to refine and will use the filters to get the sourcing pipeline that I am looking for. 

​@wmann - so nice to get to work with you on this in the Live Q&A yesterday! As you noted just above - the more generic and bare bones your JD, the less likely you will find the specific experienced candidate you want.

​@Nicole Rodney-LinkedIn  another challenge is what if you received a resume that does not have LI profile and you want to search for candidates similar to them in LI recruiter. What are ways once can do this using AI assisted job search.  Is there a way to upload a resume (if you redact personal info ) ??

​@wmann - great question! You can add a candidate to your project’s pipeline. Once you are in the pipeline, you will see a link in the upper right-hand corner: “+Add a candidate.” If you open it, you will be prompted to fill in the person’s name and contact info as well as their resume or CVC.

What I would try first is to look for keywords, skills, and job titles that the person has listed on their resume. I would use those as my search criteria. Then I would use the people who show up in my results as the starting point for my “Similar Profiles” or “Find more people like” searches. Sometimes it is about  learning what words are used by the folks doing these jobs. This approach can help.  

Now, full disclosure: I have not uploaded someone’s resume who is not a LinkedIn member. There may be other folks here who have. If so (and yes, I am asking folks reading this post who have done this - Ya, I’m really talkin’ to YOU), I hope they will let us know their experiences/successes/challenges with this. 

 

Hi ​@Nicole Rodney-LinkedIn :

Here’s one I discovered the other day exploring someone else’s issues:
Spotlights don’t have cross-referencing on them with AND/OR - so I can’t look for candidates that are BOTH Open to work AND Active Talent:

so if I try to run both spotlights simultaneously, it creates the world’s most ridiculous Venn diagram…

Hi ​@Steven Prince,

Yes, that could give you some crossover, for sure!  What I suggest is use each of those spotlight filters, one at a time. The “Open to work” spotlight is drawing from folks who have self-identified (either publicly or privately <so their current company doesn’t see they are Otw>) that they are looking for work. 

“Active talent” is giving you folks who are likely to be looking for work. They may or may not have self-identified as Open to work. These are folks who don’t always fall neatly into another category - but could be quite happy to hear from you. As the info popup notes, these are people who have either updated their profile, added their resume to their profile, or are working for a company experiencing layoffs. Very cool info to have as a recruiter.

There is no other way to easily find this type of information about any candidates on any platform, at least that I know of... It’s especially nice to be able to learn it about the ones with qualities you want. The one caveat is that we can’t easily tell which category pulled them into that spotlight for us. I suspect that folks who are applying to jobs on LI are also falling into that spotlight, although the popup doesn’t specify that.

Now keep in mind I am not an engineer, but it does make sense to me that sometimes there is crossover and sometimes not, depending on what the users in your Talent Pool search results have done on their profiles. Does that make sense? I hope it help!

Great question to ask ​@Steven Prince 

​@Nicole Rodney-LinkedIn  Thank you for explaining the Spotlight filters for Active Talent an Open to Work. I sometimes have been thinking about these two as ‘one in the same’ even though they are not the same at all. This reminded me to use these spotlight filters more often.

Wendy 

​@wmann - I am so glad you were reminded to use the Spotlight filters! Those are like your extra super-power. They are categories for the folks most likely to want to hear from a recruiter (you!). After refining searches with the candidate’s must-have criteria, check the spotlight filters and reach out to these folks!

Certain specialized jobs I hire for in IT have “very niche job titles.” For example, we sometimes hire for System Monitoring Engineer who develops automation “systems” via Python (mostly) for IT monitoring of enterprise network. They also perform some monitoring tasks and need to have skills in automation scripting via Python, server networking (especially SNMP technology experience) and preferably some Linux systems administration experience...I’ve used system engineer title with the relevant skills but often get many “bad matches.” 

Hi ​@Will B - you make an excellent point. When you use an extremely specific title (and it isn’t even one of the titles recommended by Recruiter in the Job Title search filter), you will get very few results. If you layer skills on top of that, you may get nothing.

Is all hope lost??? 

No.

A few things to try:

• Do not use a job title in your first search. Use the skills you need. Candidates may not call themselves a System Monitoring Engineer, and yet they may have the skills. This will help you uncover other job titles.
• Test different Job Titles that are recommended to you by Recruiter. These are standardized and pull in less standard titles to give you the best results. For example, Software Developer and Software Programmer both get standardized to Software Engineer. Acronyms, abbreviations, and translated titles often all roll up into the recommendations as well. This way you can get more results.
• Candidates can put whatever they want as their job title. You are asking a machine. Machines are often very specific and will give you the exact results you ask for. Ask wisely. 
• Keep in mind someone calling themselves a Software Ninja may do that on purpose in order not to show up in Recruiters’ results. They are savvy enough to understand and don’t want as much attention - another reason searching by skills is important.
• Ask Recruiter to suggest other titles or even skills - or other ways to broaden your results by typing the request in the AI-Assisted search bar and then using the applicable recommendations.
• Check for groups on LinkedIn. You can use those memberships to find candidates. I did a search for “systems monitoring’ and came up with ~1k results some with as may as 72k members. If you join those groups and start sharing relevant articles, they will get to know you. When you reach out to connect, they will be more likely to want to get to know you as you are a thought leader. Leverage that into making connections and finding candidates through their recommendations.

If you want, we can work on this search in a Live Q&A. I’m happy to strategize further with you there!

One challenge I run into with LinkedIn Recruiter is the search filters. Sometimes they work great and surface exactly the right profiles. Other times, the results feel completely off, even when I’m using the same filters I’ve used before.

Curious if anyone else sees this — and if you’ve found any tips or tricks to get more consistent results?

That is so interesting. I admit I haven’t heard this before. I wonder what has changed. (And I don’t mean the position of the stars and planets.)

Here are some questions that pop into my mind:

• Are you in the same project using the same filters or in different projects?
• Have you added or removed a job post?
• Is the location (or industry, or groups, or spotlight filter, or any other filter) different?

Any one of these variables could result in changes in the algorithm and people you are finding. So curious to hear more and hear if others have this experience too!

Depending on the search, and how specialized the job is you are searching for, there may not be other candidates who match your search criteria...thus you get candidates who are not good matches - not sure if this is the case, but I’ve had some searches where the number of qualified candidates in geographic area I’m searching are limited...in cases like this i may modify my search filters some to get better results.

I think that’s what is happening - this is really for those niche roles; I always end up tweaking my criteria even if it worked the first time 

I have had great luck pairing LinkedIn Pro w/ ChatGPT.  I use ChatGPT to recraft my job descriptions for “impact and candidate attraction”...it works fantastically.  Then, I ask ChatGPT to develop a boolean string to identify strong candidate matches.  I plug the boolean string into LinkedIn Prorecruiter, refine slightly w/ filters, and boom...it targets and returns excellent results.  I then capture the best ones into my project pipeline, (those candidate who are a better than average match, I select them and click on “find more people like” to get more targeted matches).  Then I select all the ‘standout’ candidates in my pipeline, mass email them (I have a great email intro that gets about 20% reply results) and I add the link to the job post on our website careers page (not the LinkedIn ad) into the body of the email, and this has increased my candidate pipeline dramatically...sometimes overnight.  Using this method I have hired two new employees in the last 4 weeks.  Happy Hunting!!  

Given Linkedin is a Microsoft company - I could see more integration with Copilot (vs ChatGPT)