LinkedIn Hiring Assistant 2 Rates Your Interest and Fit. Don't Ghost

Starting in November 2026, LinkedIn's recruiter AI puts two labels on every candidate it recommends: how likely you are to respond, and whether your past work matches the job. That means two habits most job seekers think cost nothing (leaving recruiter InMails unanswered and keeping a vague profile) now count against you. Here is how each signal is built and the three changes worth making before the rollout reaches your next recruiter.

What LinkedIn Hiring Assistant 2 changes in November

From November 2026, Hiring Assistant 2 attaches an interest estimate and fit signals to every candidate it recommends, pulls in people from employers' applicant tracking systems, and remembers who a recruiter passed on. Existing English-language customers get it free. LinkedIn announced it on Sept. 29, 2026 at Talent Connect, describing it this way: "before, Hiring Assistant did a search based on what you said. Now, it searches based on what you mean."

For candidates, four changes matter:

• A "candidate interest" estimate sits on each recommended profile. It predicts how receptive you are likely to be.
• "Fit signals" flag mismatches between your history and the role. ERE's launch coverage uses the example of someone who grows existing accounts being put forward for a prospecting job.
• ATS candidates join the pool. An applicant tracking system (ATS) is the software an employer uses to store applications. Hiring Assistant 2 now reads candidates in connected Greenhouse, Lever, SmartRecruiters, Zoho Recruit and Tracker accounts, with iCIMS and Workday listed as coming.
• It remembers. Who a recruiter invited before, and who they passed on, shapes what the assistant suggests next.

This matters because of what LinkedIn reports about Hiring Assistant's results. Recruiters are 4x more likely to contact candidates Hiring Assistant sourced than candidates found through traditional methods, and applicants rated as top matches are 2.6x more likely to be interviewed than lower-rated ones. Pre-screening takes a median of six minutes, and borderline candidates get sent to a screen. Whatever the AI thinks of you shapes the recruiter's shortlist before a person reads your profile.

How the LinkedIn candidate interest signal is built

The LinkedIn candidate interest signal estimates how likely you are to engage if a recruiter reaches out. Per ERE, it draws on three inputs:

1. Platform engagement: how active you are on LinkedIn.
2. Open to Work details: what your Open to Work setting says about you.
3. InMail response: how often you answer recruiter InMail.

The third input says answer, not accept, and the difference matters. Under LinkedIn's InMail policy for recruiters, a response means the candidate replies, accepts the invitation, or clicks "Not Interested." A decline counts as a response. Silence does not.

Why recruiters care about your reply rate

Recruiters care about reply rates because LinkedIn penalizes them for low ones. A Recruiter seat "must maintain an InMail response rate of at least 13% on 100 or more InMail messages sent within every 14-day assessment period." Seats that fall below that get throttled.

So every recruiter already had a reason to favor people who write back. An AI that ranks candidates by likely response turns that preference into the default ordering. You get no say in whether this happens. The only part you control is what your history tells the model.

Does ignoring InMail hurt you?

Yes. One way it hurts is documented and the other follows from LinkedIn's own description. Going quiet can switch off your Open to Work setting, which is one of the interest inputs, and silence also lowers the response history the signal reads.

The documented chain comes from LinkedIn's help center: "If we notice that you stop responding to InMails from recruiters, we'll email you to confirm that you're still open to work. If we don't receive confirmation, we'll automatically remove the #OpenToWork feature from your profile until you manually enable it again." If you miss that email, you lose the badge, and with it one of the three inputs to your interest estimate, without having done anything.

The reasoned chain is about response rate. The signal tracks how often you answer InMail, and LinkedIn's policy counts a "Not Interested" click as an answer. LinkedIn hasn't published how declines are weighted inside the interest score, so treat this as a sensible reading of the stated inputs, not a confirmed formula. Still, a decline costs almost nothing and silence has no upside.

Ignoring recruiters usually has a reason behind it. You're employed, you don't want your manager to see you replying to a competitor's recruiter, and each conversation takes an evening you don't have. So you let the messages pile up and apply privately instead, sending applications into the void. Under Hiring Assistant 2, that silent pile becomes a data point that says you won't respond.

A decline that does double duty

Don't just click "Not Interested." Use the decline to tell the model and the recruiter what would get a yes. Something like:

Thanks for reaching out. Not the right time for this one, but I'd talk about enterprise roles focused on net-new logos, ideally remote. Feel free to keep me in mind.

It counts as an answer, it costs nothing, and it gives the recruiter (and an assistant that learns from their choices) a more precise picture of you than your title does. That precision matters for the second signal.

LinkedIn recruiter AI fit signals: why your title is the problem

LinkedIn's recruiter AI fit signals flag mismatches between the work you've done and the work the job needs. For a large share of professionals, the title they hold gives the model nothing to go on. ERE's example of an account grower put forward for a prospecting role is a real risk, and it comes mostly from ambiguous titles.

In Refolk's index of US professional profiles, three sales title groups show the problem:

Source: Refolk's index, US current titles. Titles match by inclusion (so "Senior Account Manager" counts), samples contained some executive noise, and the multiples are derived. Treat the counts as directional.

The mechanism

Your title is the quickest clue to whether you hunt or farm. "SDR" says prospector. "Customer Success Manager" says grower. "Account Executive" could mean either: an AE at one company closes net-new deals all quarter, while an AE at another manages a book of renewals. In Refolk's index, Salesforce, HubSpot and Ironclad all employ US Account Executives under that one title, and the title alone doesn't say which kind of work any of them do.

When the title gives no signal, the model has to infer your type of work from your descriptions. If those read "managed a portfolio of strategic accounts," that reads as account growth, whatever you actually did. And grower titles outnumber prospector titles 8.3 to 1 in this market, so account-management language puts you squarely in the larger group.

The same ambiguity applies outside sales. "Senior Software Engineer" doesn't say whether you built the first version of something or kept a mature system running at scale. "Product Manager" doesn't say 0-to-1 or optimization.

Spell out the kind of work you actually do

To avoid a fit-mismatch flag, write in plain words which kind of work you do, on both your LinkedIn profile and your resume. Your bullets should answer the question your title leaves open, using terms a model can match against the job description.

Use the vocabulary the role itself uses:

• Prospectors: "net-new logos," "outbound pipeline sourced," "$X new ARR" (annual recurring revenue).
• Growers: "expansion," "upsell," "renewals," "NRR" (net revenue retention).
• Engineers: "built from zero to first production release" versus "scaled to X requests per second" or "reduced on-call incidents."
• Product managers: "0-to-1 launch" versus "growth and optimization of an existing product."

Two practical rules:

1. Put the type of work in your headline or first line. "Account Executive, net-new mid-market" closes the gap before the model reaches your bullets.
2. Be honest about the mix. If you did both, say so with numbers for each. A hybrid with clear figures reads better than a vague generalist.

The harder part is that the right framing depends on the posting. If you're applying to a hunter role on Monday and an expansion role on Wednesday, the same history needs different emphasis each time. Refolk takes that work off your hands: paste the posting into Refolk and you get your own resume back, rewritten to lead with the work that role needs, plus a fit score that tells you if you're stretching.

The stale resume an employer's ATS still holds

Hiring Assistant 2 re-reads candidates already sitting in connected ATSs, so the resume you sent an employer two years ago may now be your most important document there. That record stopped changing the day you applied, and the AI can flag a mismatch from it before anyone looks at your current LinkedIn profile.

The connected systems at launch are Greenhouse, Lever, SmartRecruiters, Zoho Recruit and Tracker, with iCIMS and Workday coming. If you applied through a careers site on any of those, the assistant can see you "with what your system knows about them," meaning the old application, the old title and the old bullets.

Persistent memory raises the stakes. Hiring Assistant 2 learns from who a recruiter invited and who they passed on. If an outdated record gets you passed over once, that recruiter's later searches can carry the same judgment forward.

What to do about it:

1. List the employers you've applied to in the last few years, especially ones you'd still join.
2. Check whether their candidate portal lets you update your profile or resume.
3. If it doesn't, apply to a current, relevant posting with an up-to-date resume, so the newest record reflects who you are now.
4. Make the new resume and your LinkedIn profile agree on titles, dates and type of work, so the two sources don't contradict each other.

Step 3 only works if the new resume is actually tailored, not your old PDF with a new date. Refolk builds the resume from your full history, so every posting gets a version tailored to it from the same accurate source, along with a cover letter draft.

None of this means gaming a model. Each step makes your profile say more clearly what you've done and what you want, which is what a good recruiter was always trying to work out.

FAQ

Does clicking "Not Interested" count as a reply?

Under LinkedIn's InMail policy for recruiters, yes. A response is defined as replying, accepting the invitation, or clicking "Not Interested." How Hiring Assistant 2 weights a decline inside its interest score hasn't been disclosed, but a decline counts as answering and silence doesn't.

When does Hiring Assistant 2 go live?

LinkedIn announced it on Sept. 29, 2026 at Talent Connect. Existing English-language customers get it at no extra cost starting November 2026. Recruiters you hear from after that point may be working from its rankings.

I'm an Account Executive. Will the AI think I'm a prospector or an account manager?

Your title won't settle it. In Refolk's index, 248,428 US professionals hold Account Executive titles, and the title alone doesn't say whether they win new customers or grow existing ones. Put the type of work in your headline and back it up in your bullets with "net-new logos" and new ARR, or with "expansion" and NRR, so the fit signal has something concrete to read.