LinkedIn says the next version of its recruiting agent will remember what a recruiter has hired for before and will show verification status next to every candidate. Those two changes point at the same problem: with applications up and hiring down, recruiters need help deciding whom to believe, not just whom to find.
What LinkedIn announced
On September 29, LinkedIn announced Hiring Assistant 2, the next generation of its AI agent for recruiters. The company says it adds more advanced reasoning, memory and personalization. For global customers who already use Hiring Assistant in English, the new capabilities begin to arrive automatically and at no charge in November.
LinkedIn gives the context in its own numbers. Global hiring is down 30% and applications per applicant are up 30% compared with pre-pandemic levels, according to the announcement. The company also says more than 20,000 companies now use its agentic hiring tools, including AMD, Cisco, Ochsner Health, Palo Alto Networks and Zillow, and that recruiters are 4x more likely to contact candidates sourced by Hiring Assistant than candidates found through traditional methods.
Memory moves the work from search to inference
LinkedIn’s example is specific. A recruiter used to ask the agent for a product management candidate and got one. With memory, the agent is meant to find a product management candidate close to the office, in the retail industry and with mid-level experience, “because those are all the things you’ve hired for in the past,” in the company’s words, as published in the product announcement by Dan Reid, vice president of product, and Prashanthi Padmanabhan, vice president of engineering at LinkedIn Talent Solutions.
That is a change in who writes the search. The recruiter states a role, and the system fills in the rest from hiring history, at what LinkedIn describes as role level and recruiter level. LinkedIn reports that recruiters are 27% more likely to contact a sourced candidate than they were six months ago, which the company reads as the agent getting better at finding people recruiters want to talk to.
Our read, not LinkedIn’s: any system that infers preferences from past hires will carry past patterns forward unless someone can see and change what it has learned. Customers evaluating the update will want to ask what the memory stores, whether a recruiter or an HR administrator can inspect it, and how to reset it when a role changes. AI Recruitment Tools Are Screening Out the Skills HR Wants covers how skill filters drift when nobody reviews them.
Verification moves into the shortlist
The second change addresses trust. In LinkedIn’s research, run by Censuswide for LinkedIn in August 2026 among 500 US recruiters and 500 US executives at companies with 200 or more employees, 64% of talent acquisition professionals say it is getting harder to know who to trust and what is real when assessing candidates. More than one-third say their top challenge is not knowing whether a candidate or their skills are real. Because the research was commissioned by the vendor, treat the figures as LinkedIn’s framing of the market.
Hiring Assistant 2 puts more signals in one view: whether the candidate is verified, applicant tracking system insights, pre-screening responses and relevant work experience. LinkedIn says 115 million members are verified, and that verifications will now appear inside the applicant tracking systems of partners including iCIMS, JazzHR, Jobvite, Lever, SmartRecruiters and Workable.
Skills evidence is the other half. LinkedIn is expanding connected apps, which let members show work done in outside tools on their profile, to partners including Chess.com, ElevenLabs and Medium. A recruiter searching for someone who has built voice agents, LinkedIn says, would see members with validated capabilities from ElevenLabs.
The agent can also manage a whole talent pool, including candidates who applied outside LinkedIn, once a recruiter connects an applicant tracking system. LinkedIn lists Greenhouse, Lever, SmartRecruiters, Zoho Recruit and Tracker as available now, with iCIMS and Workday coming soon. Voice is being added to pre-screening, and LinkedIn says pre-screening completes in a median of six minutes.
What the customers say
LinkedIn’s announcement carries comments from customers and partners. Scott McGuckin, VP of Global Talent Acquisition at Cisco, said: “It helps us move faster while keeping human judgment and trust at the center of Cisco’s recruiting.” Roz Harris, VP of Talent at Zillow, said that during a pilot, sourcing efficiency improved by roughly a third.
The efficiency figures in the announcement come from LinkedIn: recruiters find qualified matches while reviewing 83% fewer profiles, find interview-quality applicants 33% faster, and applicants rated as top matches are 2.6x more likely to be interviewed than lower-rated ones. Those are vendor-reported measures of the vendor’s own ranking. The last one shows that recruiters follow the ranking, which says nothing yet about whether the ranking is right.
What it means for the HR leader
Recruiting agents are moving from a search box to a system with memory and a say in the shortlist, and the buying questions change with it.
- Ask about memory before the November update lands. Find out where hiring preferences are stored, who can read them, and how a team clears them.
- Decide how verification will be used. A verified badge is a signal for the recruiter. If it quietly becomes a filter, candidates who are not on the platform or have not verified are screened out by default. Write down which one you intend.
- Check your applicant tracking system against the partner lists. The integrations named above are the ones LinkedIn has announced, and some are still marked coming soon.
- Measure your own funnel. Compare interview rates and time to first response before and after, instead of relying on the vendor’s medians.
- Keep the audit trail. Employers answer for hiring outcomes whichever tool produced the shortlist, as our Q4 opinion on inventorying hiring AI argued.
Source: LinkedIn Newsroom
