Six years ago, I wrote about the four habits of a successful sourcer. Back then it was about curiosity: Googling everything, reverse-engineering URLs, reading GitHub code. Those habits still matter. But if you asked me today what separates a great sourcer from an average one, I’d add one thing: the ability to change how you approach the work itself.
AI hasn’t just given sourcers new tools. It has changed the logic of sourcing. The sourcers who are struggling now aren’t struggling because the tools are hard, they’re struggling because the tools demand a fundamentally different way of thinking.
Here are four mindset shifts my team went through this past year. Each was uncomfortable at first. Each made us stronger.
Stop narrowing your search. Start casting a wide net.
Sourcers are trained to be precise. You open LinkedIn, build a tight Boolean string, add every filter, and expect the most relevant profiles first. The tighter the search, the better the result.
Now try that approach in Clay, a lead-generation and data-enrichment platform that works brilliantly for sourcing, but with completely different logic. Come expecting a narrow, super-filtered search like LinkedIn, and you’ll get weak results.
The right approach is the opposite. In Clay you cast a wide net first, say, all Frontend Developers in a given country. That gives you a large pool. Then you rank and filter inside the tool: remove juniors, exclude managers and team leads depending on the role. The list narrows, but the narrowing happens after the initial pool, not during it.
We often take the ranked list from Clay and load it into Claude, where AI scores each candidate against the job requirements (A-tier, B-tier, C-tier) with written reasoning. The sourcer reviews the AI’s logic, overrides where they disagree, and gets a clean prioritized list in a fraction of the time it used to take.
Takeaway: your job is no longer finding the needle in the haystack. It’s building the haystack smartly and letting AI help you sort it.
You no longer need to read every profile.
For most sourcers, reading profiles is the heart of the job. You pull a list, open each profile, scan experience, check whether the person worked at a product company, knows a specific technology, whether their career trajectory fits the role. It’s slow, manual, and sourcers take pride in doing it well.
So when I told my team that AI could do this step for them, the first reaction wasn’t enthusiasm. It was resistance.
The new workflow: you have 300 candidates from Clay. Instead of opening each profile, load the list into Claude Cowork with the job description. Claude Cowork checks each candidate: product-company experience? React and TypeScript? Right team size? Hands-on or management? Output: a ranked list with explanations for each score.
Is it perfect? No. Sometimes AI misreads a profile or weighs a criterion differently. But reviewing and correcting AI output takes 20 minutes. The same work manually takes an entire day.
One of our sourcers said it best: “I used to spend all morning reading profiles. Now I spend 20 minutes checking what Claude Cowork flagged, and the rest of the morning actually talking to the best candidates.” That’s the shift — from a person who reads profiles to a person who verifies AI judgment and has real conversations.
Takeaway: your value isn’t in reading profiles. It’s in what you do with that information next.
You no longer need to write every message by hand.
Every sourcer has outreach templates. And every sourcer knows that “I came across your profile on LinkedIn” doesn’t work anymore. Candidates spot templated messages instantly. Personalization matters. But personalizing 300 messages a week by hand is a full-time job on its own.
This is where the combination of Clay and a multi-channel outreach tool like Lemlist changes the game. Clay enriches each candidate profile with data far beyond what LinkedIn shows: company details, size, tech stack, recent job changes. When you feed that enriched data into Claude along with the job description and your outreach tone guidelines, it generates personalized sequences for each candidate not one message, but a full sequence: LinkedIn connection request, first message, follow-up email, second follow-up with a different angle.
Because each message draws on real candidate data, the personalization is genuine. But the sourcer still owns quality — AI generates the draft, you own the final version. And what truly matters is the conversation after a candidate replies: reading tone, understanding what they really want, building trust, knowing when to push and when to step back. No AI can do that.
Example of a Lemlist campaign for full-stack engineer position
Takeaway: your craft isn’t in writing the first message. It’s in the conversation that follows.
You’re no longer a sourcer. You’re the manager of an AI engine.
This is the biggest shift and the hardest to accept.
In the traditional model, a sourcer’s week looks like this: search for candidates, read profiles, write messages, send messages, track responses in a spreadsheet, compile a report on Friday. Every step is manual. The sourcer is the operator of every task.
In the AI model, the week looks completely different. Clay pulls candidates. Claude ranks them. Outreach sequences are generated and loaded into Lemlist. Responses are tracked automatically in Airtable. The weekly report is generated by AI from data already in the system.
So what does the sourcer do? Builds and manages the entire system. Sets up Clay filters based on the ideal candidate profile. Reviews AI rankings and catches errors. Approves outreach sequences. Leads live conversations with interested candidates. Reads analytics and makes fast decisions: is the low open rate a subject-line problem or a deliverability issue? Are we hitting the right seniority, or do we need to change the filters?
One of our team leads said something that stuck with me: “It used to feel like my job was sending messages. Now it feels like my job is making sure the right candidates hear the right story at the right time.”
Takeaway: you’re no longer the person who executes tasks. You’re the person who makes sure the system works and steps in where only a human can make the difference.
Final thought
When I wrote about sourcer habits six years ago, the core message was curiosity. That message hasn’t changed. AI has made curiosity even more important, not less.
But curiosity alone is no longer enough. You need willingness to let go of how you’ve always worked. A sourcer who insists on manually reading every profile isn’t thorough — they’re slow. A sourcer who refuses to trust AI ranking isn’t cautious — they’re leaving opportunities on the table.
The best sourcers I work with today made these shifts early. Not because they’re more technical, but because they were willing to feel uncomfortable. They tried the wide net when it felt wrong. Let AI review profiles when it felt risky. Launched AI-generated outreach when it felt impersonal. And then saw the results.
Tools will change. They always do. But these four mindset shifts aren’t about any specific tool. They’re about how a sourcer thinks about their role. And that is what will separate the sourcers who thrive from those who get left behind.
