AI for LinkedIn recruiting
Two constraints shape everything here and are missing from the generic advice: a finite monthly InMail allowance, and the requirement that a shortlist be defensible. These prompts score only what a profile can evidence, and they refuse to infer protected characteristics or their proxies. The candidate message itself is the same craft as the sales prompts, written out in full there and sized to the same weekly ceiling.
Map before you search. A week of sourcing from the wrong three companies is the most common way a role goes unfilled for a quarter.
Withholding the company name and the salary from the first message. It halves the reply rate and gains you nothing — candidates read the silence as the role being worse than the message implies.
3 for this work
- Prompt
Candidate Research & Shortlist
A ranked shortlist with evidence per requirement, the gaps named, and a screening question per candidate.
- Prompt
Recruiter Outreach
A subject line and message aimed at one candidate, with the details that decide whether a passive candidate replies included rather than withheld.
- Prompt
Talent Map
A named list of source companies, the titles the role is called at each, and the search strings that find them.
Attach these once instead of retyping
A prompt is consumed once; a context file is read on every turn for months. These are the ones this work depends on.
- Context file
RECRUITER.md
Sourcing and outreach that stays inside your InMail budget and does not infer things it should not.
These are steps in a longer job
From open role to a shortlist worth defending runs 4 steps end to end, and several of them are on this page. Each step names what it hands the next one, which is the part a list of prompts cannot show.
Every AI feature in Recruiter sits on the tier LinkedIn will not price, and Recruiter Lite unlocks none of them
This is the fact the generic advice leaves out. LinkedIn’s own four-tier comparison marks Recruiter Lite as having no AI features and no Hiring Assistant, and the tier above it carries no published list price at all. Hiring Assistant is then an add-on on top of that — a second unpublished number, not a feature of the seat.
| What LinkedIn calls it | What it does | Availability | What unlocks it | What that costs |
|---|---|---|---|---|
| AI Search | You describe the candidate you want, or paste a job description, and Recruiter turns that into filters instead of you setting them. | Gradual rollout | Recruiter Corporate | Quote only — LinkedIn prices Recruiter Corporate through a representative |
| AI Messages | It drafts a personalised InMail to a candidate, and will do it for a batch at once. | Gradual rollout | Recruiter Corporate | Quote only — LinkedIn prices Recruiter Corporate through a representative |
| AI-Assisted Job Posts | It generates a job post, or reformats one you already have. | Gradual rollout | Recruiter Corporate | Quote only — LinkedIn prices Recruiter Corporate through a representative |
| AI-Assisted Applicant Management | On the Talent pool page you narrow a large applicant pool by typing the criteria instead of filtering by hand. | Gradual rollout | Recruiter Corporate | Quote only — LinkedIn prices Recruiter Corporate through a representative |
| Hiring Assistant | It sources candidates into a pipeline, evaluates applicants against the role, writes fit summaries and drafts outreach. LinkedIn launched it to charter customers in October 2024 and made it globally available in English by the end of September 2025. | Gradual rollout | Recruiter Corporate | Quote only for Recruiter Corporate, plus an add-on LinkedIn publishes no figure for |
Feature names are LinkedIn's own, read from its help centre and product pages on 20 September 2026; the prices come from this site's plan records and carry their own read dates per product family. Availability is LinkedIn's word, not a measurement — a gradual rollout means some accounts have it and LinkedIn does not publish which. The full records, with the ceilings and what LinkedIn declines to disclose, are on the LinkedIn AI features page.
10 tools name Recruiter, and every one of them reaches it through a session rather than an ATS route
LinkedIn does have a sanctioned integration path for recruiting — the ATS partner programme — and nothing in this group is on it. What they are is contact-data extensions that read a Recruiter Lite page, and outreach automation pointed at candidates. The difference decides who carries the consequence.
It reaches whatever the page in front of you shows, and nothing while the browser is closed. LinkedIn’s Prohibited Software and Extensions policy names extensions that automate the site, and enforcement lands on the member account.
The same reach as the extension class and the same clause against it, with one real difference: the session stays on your machine rather than in a vendor’s datacentre. The computer has to be awake for anything to run, and moving it to a rented server to fix that puts the traffic back on a datacentre IP.
The widest reach of any class and the highest exposure. It runs without you, the credential sits with somebody else, and every action is attributable to your account — which is also the thing LinkedIn can restrict.
| Tool | How it reaches LinkedIn | What it does with your account | Account exposure |
|---|---|---|---|
| Cognism | Chrome extension | Extracts automatically | Medium |
| Linked Helper 2 | Desktop app | Sends under your name | High |
| Dripify | Vendor cloud | Sends under your name | High |
One tool per route, alphabetically first within each, so nothing here reads as a pick — and exposure is derived from the route and whether the tool sends, not typed next to the vendor it flatters. All 10 in this group, filterable, are in the LinkedIn Recruiting AI directory; every record was read from the vendor's own pages on 20 September 2026.
1 of the 13 MCP servers reaches Recruiter at all, and only partly
The rest read public profiles, company pages and job postings, which is genuinely useful for mapping a market — and does not touch a seat. The servers below are the ones in that group, ordered by what they expose.
A hosted, paid server that returns structured LinkedIn data — profiles, people search, companies, posts, jobs, the Ad Library — plus a dozen other sources, with server-side filtering and export. No LinkedIn login goes anywhere near it, and it cannot write to LinkedIn at all.
A general web-access server with five LinkedIn extractors: hand it a profile, company, job, post or people-search URL and it returns structured JSON. It holds no LinkedIn credential, so there is no account to restrict.
Research and record-keeping for manual outreach: it briefs you on a person or a company, mines a post's comment thread, and keeps a local ledger of what you sent and what replied. It has no send path at all.
Reads public LinkedIn company and person pages, and a company's public posts when LinkedIn serves them signed out. It never logs in, so it puts no account at risk — and cannot see anything behind the login.
4 of the 10 servers filed under profile and company data, ordered by account exposure, lowest first. That is an ordering, not a recommendation: LinkedIn publishes no MCP server of its own and its User Agreement prohibits automated access by any means, so the directory puts that clause beside each record rather than ranking them. Read 20 September 2026.
What a sourcing prompt cannot do, and must not be asked to
Stated here rather than left for a reader to discover, because the gap between what a prompt does and what people expect it to do is where the time goes.
The allowance is monthly, credits expire after 90 days, and they accrue to at most three times the monthly figure. A prompt that plans 60 candidate messages on a 30-credit month has planned a month you do not have — which is why the outreach prompts here ask for the allowance first. Allowances per plan, and what a credit is worth.
A shortlist prompt scores evidence, not truth. It marks a requirement met only where the profile itself shows it, and marks the rest not evidenced rather than inferring it — which is also the only form a screening decision can be defended in.
Not age, not nationality, not parental status, and not the proxies — graduation year, name origin, career gaps. The prompts here refuse those fields explicitly, because a model asked to rank candidates will otherwise use them without being told to.
A ranked shortlist with evidence per requirement is an input to a human decision and is worded as one. The output names what it could not establish, which is the part a confident summary drops.
The questions people arrive with, answered directly
Does LinkedIn Recruiter have AI, and which tier do I need?
It has AI Search, AI Messages, AI-Assisted Job Posts, AI-Assisted Applicant Management and AI insights in the InMail Report, all in gradual rollout. LinkedIn’s own comparison marks Recruiter Lite — $170 a month, read on 10 September 2026 — as having none of them. The tier that does is Recruiter Corporate, and LinkedIn publishes no list price for it.
How much does LinkedIn Hiring Assistant cost?
LinkedIn publishes no price: not a list price, not a per-credit price, not a tier price. It is an add-on on top of LinkedIn Recruiter, which is itself quote-only, so the answer is two unpublished numbers stacked. What LinkedIn does publish is 6 credit tiers describing how many candidates it may source and evaluate per licence per month, and those are worth taking into the sales call.
Can AI source candidates on LinkedIn for me?
Not from a chat window — LinkedIn exposes no people-search permission to a self-serve app and blocks assistant crawlers. Inside Recruiter, AI Search turns a description or a pasted job description into filters, on the quote-only tier. Outside it, 10 third-party tools point at candidates and all of them reach LinkedIn through a session, with the account carrying the consequence.
Is it safe to use AI to screen LinkedIn candidates?
Safe depends on what the model is allowed to score. The defensible form is evidence-based: each requirement marked met only where the profile shows it, and marked not evidenced otherwise, with no inference of protected characteristics or their proxies. The prompts here are written that way and say what they could not establish, which is exactly the part a confident AI summary removes.
Everything for recruiting, in one folder
The Starter Pack bundles the twelve context files and the twelve most-used prompts as a zip — including the ones on this page. Free, and the ungated link sits under the form.