AI Candidate Sourcing on LinkedIn: Step-by-Step

Why Traditional LinkedIn Sourcing is Broken

Here is how most recruiters source on LinkedIn: type a job title into the search bar, scroll through results, click on profiles that look promising, send a generic InMail, and hope for a reply.

The response rate for this approach? Between 5% and 15%. That means for every 100 InMails you send, 85 to 95 people ignore you completely.

AI does not fix LinkedIn sourcing by magic. It fixes it by helping you find better candidates, write better messages, and target the people most likely to respond. The result is response rates of 25% to 35%, which means the same effort produces twice as many conversations.

Step 1: Build Your Ideal Candidate Profile with AI

Before you search for anyone, define exactly who you are looking for. Use this prompt:

"I need to source candidates for a [job title] role at a [company type] in [location]. The salary is [range]. Key requirements are [list]. Based on these requirements, create an ideal candidate profile including: likely current job titles (5 variations), companies they might work at, skills that should appear on their profile, certifications or qualifications to look for, and keywords that indicate seniority level."

This gives you a sourcing blueprint. Instead of searching for one job title, you now have 5 variations to search. Instead of guessing which companies to target, you have a list. This alone doubles your candidate pool.

Step 2: Generate Boolean Strings with AI

Take the ideal candidate profile from Step 1 and use it to build search strings:

"Based on this ideal candidate profile, generate 5 Boolean search strings I can use on LinkedIn Recruiter. Include variations of job titles, key skills, and relevant qualifications. Each string should target a slightly different segment of the talent pool."

Run each string separately and review the results. The first string might return 500 profiles. The third string might return 80 highly targeted profiles that you would never have found with a basic search.

Step 3: Prioritise Candidates with AI Analysis

Once you have a list of potential candidates, use AI to prioritise who to contact first.

Copy the key details from each LinkedIn profile (headline, current role, experience summary) and paste them into Claude or ChatGPT with this prompt:

"Here are 10 candidate profiles for a [role title]. Rank them from most to least suitable based on: relevance of current role, years of experience, industry match, and likelihood of being open to a new opportunity. For each candidate, give a 1-sentence rationale for the ranking."

This saves you from spending 20 minutes deciding who to contact first. The AI surfaces the strongest candidates instantly.

Step 4: Write Personalised Outreach with AI

This is where AI makes the biggest difference. Instead of sending the same template to everyone, use AI to personalise each message based on the candidate profile.

"Write a LinkedIn InMail to [candidate name] who is currently a [current role] at [current company]. The role I am recruiting for is [role title] at [company name] in [location] with a salary of [range]. Personalise the message based on their experience at [current company] and their skills in [relevant skills]. Keep it under 150 words. Sound like a real person, not a recruiter using a template."

The personalisation does not need to be deep. Mentioning their current company, a specific skill, or a recent career move is enough to show you have actually looked at their profile. That alone increases response rates by 40% compared to generic templates.

Step 5: Automate Follow-Ups

Most candidates do not respond to the first message. That is normal. The placement often comes from the second or third touchpoint.

Use AI to generate a 3-message follow-up sequence:

• Message 1 (Day 0): Initial personalised outreach
• Message 2 (Day 4): Brief follow-up adding one new piece of information (salary detail, team size, growth opportunity)
• Message 3 (Day 8): Final nudge, slightly different angle ("Even if the timing is not right, happy to have an informal chat about the market")

Generate all three messages at once so they feel like a natural conversation, not disconnected automated blasts.

Measuring Your AI Sourcing Performance

Track these metrics to know if your AI-assisted sourcing is working:

• InMail response rate: target 25%+ (industry average is 10-18%)
• Profiles viewed to shortlist ratio: target 1 in 5 (you should shortlist 20% of profiles you review)
• Time to first shortlist: target under 4 hours from receiving the job spec
• Source quality: what percentage of AI-sourced candidates make it past first interview? Target 50%+

If your numbers are below these targets, the issue is usually in your search strings (too broad) or your outreach (too generic). Go back to Steps 2 and 4 and refine.

The Bottom Line

AI does not replace your sourcing instinct. It amplifies it. The 5-step process above takes a task that used to consume an entire day and compresses it into 2 to 3 hours of focused, high-quality work.

The recruiters who master AI-assisted LinkedIn sourcing in 2026 will consistently outperform those who are still scrolling and hoping. Start with Step 1 today and see the difference for yourself.

For the complete LinkedIn sourcing toolkit with prompts, templates, and Boolean string libraries, get the Pro Playbooks guide.