Refolk
10 min readMarkdown

LinkedIn's Hiring Assistant 2 Reads Your Profile, Then Audits Your Resume

LinkedIn's Hiring Assistant 2 scores your profile, screening answers and ATS history before a recruiter opens your resume. Here is what to rewrite first.

If the company you're applying to uses LinkedIn Recruiter, an AI agent may already have formed a view of you from your profile, your screening answers and the employer's ATS records before anyone opens your resume. Your tailored resume shows up after that picture exists, and it gets checked against it.

Hiring Assistant 2 starts rolling out in November 2026. Recruiters who already use Hiring Assistant in English get it automatically and at no extra cost.

What does LinkedIn Hiring Assistant 2 actually read?

Hiring Assistant 2 scores three things against the qualifications a recruiter set at intake: your LinkedIn profile, any resume you uploaded and your screening-question answers. It then combines them with ATS records, work history and identity validation to flag inconsistencies. It also reportedly reads your posts and comments.

Hiring Assistant is LinkedIn's recruiter AI agent. It sits inside LinkedIn Recruiter and sources, ranks and pre-screens candidates. LinkedIn announced version 2 at Talent Connect in New York City. Coverage gives the date as either September 29 or October 1, 2026. Launch coverage reports that agentic hiring has passed $450M in annualized revenue, so this is a core product, not a side experiment.

InputWhat the agent does with itBasis
LinkedIn profileScored against intake qualificationsLaunch coverage
Uploaded resumeScored alongside the profileLaunch coverage
Screening answersScored, added to the candidate portraitLaunch coverage
ATS recordsPast applications, stages, consistency checksGreenhouse, Lever, SmartRecruiters, Zoho Recruit, Tracker at launch; iCIMS and Workday "coming soon"
Identity validationCombined with the above to flag inconsistenciesLaunch coverage
Posts and commentsReportedly scannedReported in Calcalist, not in LinkedIn docs

Recruiter Lite includes none of these AI features. If a recruiter is on Lite, none of what follows applies to them.

How recruiters find candidates on LinkedIn now

On full LinkedIn Recruiter, the agent runs five kinds of search side by side and hands the recruiter a ranked list. Whether you make that list depends on your profile, not on how good your resume is.

The reason comes from a technical write-up of the system: plain keyword search returned zero results for nearly half of recruiter queries. So the agent runs all of these at once:

  1. Faceted search. Hard filters on fields like location, title and seniority.
  2. Boolean search. Literal AND/OR/NOT matching on words.
  3. Semantic retrieval via MUSE. Matching on meaning rather than exact words.
  4. Recruiter-personalized recommendations. Ranking shaped by this recruiter's past choices.
  5. Lookalike search. Candidates who resemble people the recruiter already liked.

Dan Reid, LinkedIn's VP of product management, summed up the design goal: "Do what I meant, not what I said."

LinkedIn's own numbers for version 1 show what happens next. Recruiters were 4x more likely to contact candidates the agent sourced. They reviewed 83% fewer profiles. Pre-screening took a median of 6 minutes, against a typical 3.5 hours. When a recruiter looks at 83% fewer profiles, the cut happens before any person is involved. If the agent leaves you off the shortlist, nobody reads your bullets.

This also explains why so many applications seem to go nowhere. The agent doesn't only rank people who applied. It also sources people who never applied, based on their profiles. With a thin profile, you can keep sending carefully tailored applications into a queue while the recruiter's shortlist fills up with people whose profiles were already strong.

Why your job title needs two names

Put your local title and the market's common title in both your headline and your experience entries. The agent's semantic layer understands synonyms, but its faceted and boolean filters still match words literally, and they all run in the same search.

Refolk's index of professional profiles shows how big the vocabulary gap can be, using recruiters themselves as the example. In the US, 123,399 professionals hold Recruiter, Technical Recruiter, Talent Acquisition Partner or Sourcer titles. Technical Recruiter was the most common in a small 25-profile sample (9 of 25). The UK mostly uses different words for the same job.

Market / title setProfiles in Refolk's index
US: Recruiter, Technical Recruiter, TA Partner, Sourcer123,399
UK: the same US-style titles9,403
UK: Recruitment Consultant, Resourcer, TA Specialist, Recruitment Partner32,735
UK combined (small overlap possible)~42,138

A US-title search makes the US pool look 13.1x the size of the UK's. Counting the UK's local titles brings that down to 2.9x. The ratio isn't really about market size. It shows how many people a literal search misses.

22%
Share of the UK recruiter pool a US-title search finds

Derived from Refolk's index: 9,403 US-style titles out of roughly 42,138 UK recruiters.

Candidates have the same problem: "Software Developer" vs "Software Engineer," "Analyst" vs "Associate," "Resourcer" vs "Sourcer." Semantic retrieval will probably connect them. A faceted title filter won't. And if the recruiter adds a boolean clause, it's checking exact strings. Don't count on the AI to translate for you. Write it as "Recruitment Consultant (Technical Recruiter)" or "Software Developer, Backend Engineer" so every layer finds you. The same goes for your headline: an official title, the common title and a few core skills will get through a filter. A slogan won't.

LinkedIn profile vs resume: tailor emphasis, never facts

Your tailored resume can change which parts of your history it emphasizes. It can't change titles, dates or employers, because Hiring Assistant 2 is built to flag exactly those differences. Launch coverage says the agent combines identity validations with ATS data, pre-screening answers and work history to "build a fuller candidate portrait and flag inconsistencies."

That's new. Before, a recruiter might notice a mismatch between your resume and profile if they had both open in tabs. Now the software is designed to do that comparison for them.

ElementTailor per posting?Why
Job titlesNo, keep official titleMismatch can be flagged against profile and ATS
Employment datesNoEasiest inconsistency to detect
EmployersNoChecked against work history and ATS records
Bullet order and wordingYesChanges emphasis without changing facts
Summary and skills orderYesMirrors the posting's priorities

One case to watch is a title you've translated for the market. If your profile says "Software Developer, Backend Engineer," your resume can use the same pairing. It shouldn't drop your official title and keep only the friendlier one.

Refolk is built to keep you on the right side of that line. I write your resume from your own history, then tailor it to each posting by reordering and rewording things that are already true. Titles, dates and employers stay matched to your profile, so the agent has no gap to flag.

The agent compares you with the recruiter's past hires

Hiring Assistant 2 has memory. It ranks you against the people this recruiter has invited, passed on and hired before, not only against the job posting. One description of the launch puts it this way: "who you invited before and who you passed on shapes what it suggests next."

LinkedIn's own example is a recruiter whose past hires were close to the office, from the retail industry and at mid-level experience. The agent learns that pattern and looks for more of it. That makes three fields as important as your skills list:

  • Location. If your profile shows a city two states away and the recruiter's hires were local, you can rank lower before anyone asks whether you'd relocate. Put your actual metro, and say whether you're open to relocating.
  • Industry. Industry is one of the traits in LinkedIn's own example. If you're changing industries, your About section and bullets should use the target industry's words, not only the old one's.
  • Seniority wording. "Lead," "Senior" and "Manager" mean different things at different companies. Make scope explicit (team size, budget, reports) so the agent doesn't put you in the wrong band.

Screening answers are now part of your profile

Your screening-question answers are now scored alongside your profile and resume, and they feed the candidate portrait the agent builds of you at that employer. Treat them like resume bullets, not form fields.

There's a catch. A polished, AI-written answer may get discounted, because 64% of US talent acquisition professionals say it has become harder to know who to trust. A generic answer to "Describe your experience with X" reads as noise. A specific answer with a project, a number and a date reads as signal, especially when the same project appears in your experience entries.

  • Reuse the exact project names and figures from your profile.
  • Answer the qualification that was asked, not a nearby one.
  • Name a gap honestly if you have one, and say how you'd close it.

Knowing where your gap is before the form asks is half the battle. Refolk scores how well you fit each posting against your own history, so you can see which requirement is weak and write the honest version of that answer first, instead of improvising under a character limit.

Verification, connected apps and comments are the scarce signals

Proof of identity and proof of skill are rare on LinkedIn, so they stand out. Getting verified and connecting a skills app are among the few profile changes with a published payoff.

LinkedIn reports 115 million verified members. One analyst estimates that's about 9% of accounts registered globally. If fewer than one in ten profiles carries verification, an agent built to flag inconsistencies has less to question about the ones that do. Connected apps from partners such as Base44, Fiverr, Duolingo and HubSpot attach evidence of skills to your profile.

2.9x
More recruiter messages for members with a connected app

LinkedIn's own figure ("up to"), reported in launch coverage.

The posts-and-comments claim is less certain. It comes from Thieberg, interviewed by Calcalist, not from LinkedIn's documentation: "What is new is that AI scans not only what is written in the profile, but also posts and comments." Thieberg adds that "candidates are not making enough use of comments on the platform." Treat it as plausible and cheap to act on. A few substantive comments on posts in your field add searchable text in your own voice. Avoid hot takes you wouldn't want summarized for a hiring manager.

Reapplying to the same company is no longer a fresh start

Once an employer connects its ATS, Hiring Assistant 2 can see your earlier applications and the stages they reached, including applications you submitted outside LinkedIn. A second application to the same company gets read in the context of the first.

At launch, the connected systems are Greenhouse, Lever, SmartRecruiters, Zoho Recruit and Tracker. iCIMS and Workday are "coming soon," so for now, applications to companies on those systems aren't part of the agent's view.

What this means in practice:

  • Keep facts consistent across attempts. If you applied last year with one set of titles and dates and reapply with different ones, the agent has both records.
  • Change the substance, not the story. If you got rejected at the screen, the new application should show a real change: a new skill, a shipped project, more scope.
  • Find out which ATS the company uses before you apply. That tells you whether your history comes along with you.

FAQ

Does LinkedIn Hiring Assistant 2 read my resume?

Yes, if you upload one. It scores uploaded resumes alongside your profile and screening answers against the recruiter's intake qualifications. But the profile is what it sees first, including for people who never applied, and the resume is checked against it. Any difference in titles, dates or employers can be flagged as an inconsistency.

When does Hiring Assistant 2 reach recruiters?

LinkedIn announced it at Talent Connect in New York in late September or early October 2026, depending on the source. Existing customers who use Hiring Assistant in English get the new features automatically, free, starting November 2026. Recruiter Lite users don't get any of the AI features.

Should my LinkedIn title match my resume exactly?

Your official title should match exactly on both. Next to it you can add the market's common title, for example "Recruitment Consultant (Technical Recruiter)," on both documents. That gets you through literal filters without creating a mismatch. Leave the facts alone and tailor bullet order and wording instead.

Does the agent really read my LinkedIn comments?

That's reported, not confirmed. The claim comes from an interview in Calcalist, not from LinkedIn's documentation. Commenting in your field costs little and adds searchable text about what you know, so it's worth doing. Write only what you'd be comfortable seeing summarized for a recruiter.

The data behind this

Put this to work

Paste your career in once. Every application after that is written for you.

Drop a resume or a LinkedIn URL. I rank the live openings against it, rewrite the resume and write a cover letter for the best of them, and fill in the employer's form when you press the button. You read, you decide what goes out.

  1. 01Drop your resume

    A PDF or a LinkedIn URL. About a minute, once.

  2. 02I rank the openings

    Every weekday morning, the live catalog scored against your history. Up to 20 worth your time, not two hundred links.

  3. 03Each one is written up

    Resume rewritten for the posting, a cover letter, a fit score. Press send, or let me fill in the form.

  • New matches ranked and written before you are up.
  • Every bullet stays inside what your history supports. Nothing invented.
  • Queued, submitted, interviewing, offer: one screen, not a spreadsheet.

Free, no card. Up to 25 applications sent a week, and nothing is sent until you say so.

Keep reading