The Boolean Strings Recruiters Still Use: Working Examples, and What Stopped Working

Boolean search still rests on five building blocks: AND, OR, NOT, quotation marks and parentheses. SeekOut, Workable and AIHR all teach the same set. What changed is where the strings work: Google X-ray searches of LinkedIn lost access to key profile fields in early 2024, while platform search, GitHub and niche sites still respond well.

No public source documents what "top headhunters" type day to day. This article covers the strings that recruiting vendors and sourcing trainers publish, with each one attributed. Details below are as of 15 October 2026.

Which operators do most of the work in a Boolean string?

Five operators cover nearly every candidate search: AND narrows, OR widens, NOT excludes, quotation marks lock a phrase, and parentheses set the order of logic. The three guides above describe the same five.

Parentheses change the meaning of a string. SeekOut shows that sales AND (retail OR consumer) requires "sales" in every result, while (sales AND retail) OR consumer does not. Workable makes the same point with (developer OR designer) AND Java, which demands Java from both groups.

Quotation marks matter just as much. Without them, a search for project management can return pages that contain "project" and "management" anywhere, according to SeekOut.

How do you build a string that returns real candidates?

Start with job titles joined by OR, add required skills joined by AND, add qualifiers that prove the experience, then trim noise with NOT. This is the formula Pin, a sourcing vendor, uses for its template strings.

The title group is where most strings fail. Workable points out that searching only "Web developer" misses people whose titles read "Software developer" or "Web programmer", so it recommends combining AND and OR to cover the variants. Pin's cheat sheet says the same: brainstorm several title variations for every role.

A practical build order:

1. Write the three to five titles a candidate might use, joined by OR inside parentheses.
2. Add one or two required skills or tools, joined by AND.
3. Add a qualifier that signals seniority or results.
4. Add exclusions last, and check what they remove.
5. Run the string on the platform it was written for.

What do published role strings look like?

Vendors publish templates by role, and they follow the title, skill, qualifier, exclusion pattern. The strings below are copied from vendor guides as examples of structure, not as tested or guaranteed results.

Registered nurse, for a LinkedIn keyword field (from Pin's cheat sheet):

("registered nurse" OR "RN" OR "BSN") AND ("ICU" OR "emergency" OR "critical care" OR "med-surg") AND ("licensed" OR "certification") NOT "student" NOT "aide" NOT "assistant"

VP or C-suite, for Google X-ray (from Pin's advanced strings guide):

site:linkedin.com/in/ ("vice president" OR VP OR "chief" OR "C-suite" OR "general manager" OR "managing director") ("P&L" OR "revenue growth" OR "board of directors" OR "IPO" OR "M&A" OR "private equity" OR "transformation") (SaaS OR "technology" OR "software" OR "digital") -recruiter -consultant -advisor -"open to work"

Field search inside a sourcing platform (from SeekOut's help centre, the vendor's own syntax):

cur_title:finance AND degrees:bachelors AND majors:(finance OR accounting) AND NOT cur_company:SeekOut

That last string uses fields that only SeekOut documents. Field names such as cur_title: and past_titles: are not portable to Google or LinkedIn.

Pin's executive string uses the site: operator on LinkedIn, which runs into the problem covered below. Treat the business-impact terms in it (P&L, M&A, IPO) as the reusable idea, and the site:linkedin.com/in/ prefix as the fragile part.

Which Google operators can you rely on for X-ray searches?

Google Search Help documents quotation marks, site:, the minus sign, before:, after: and filetype:. It also says to leave no space between an operator and its term, so site:nytimes.com works and site: nytimes.com does not (Google Search Help).

X-ray search is the recruiting use of site:. Pin gives the pattern as site:[domain/path] "keyword 1" "keyword 2" "location" -exclusions (Pin). Workable adds intitle: and inurl: for finding resume pages and filetype: for resumes or portfolios.

One example from Pin targets GitHub profiles while filtering out project pages:

site:github.com "TypeScript" "React" "Seattle" -intitle:issues -intitle:pull

Two cautions apply. Google does not recognise NOT, so use the minus sign, as both Workable and AIHR note. And the cache: operator is gone: Search Engine Journal reported on 24 September 2024 that Google removed its documentation because the operator no longer works.

Did LinkedIn X-ray search stop working?

Substantially, according to the 2024 reports. Three separate publishers reported in early 2024 that Google lost access to key LinkedIn profile fields. Boolean Strings wrote on 5 February 2024 that Google no longer finds people by Headline, About, Experience and Education. PCRecruiter and the Full Stack Recruiter newsletter reported the same shift.

The reports differ on how total the loss is. PCRecruiter says X-ray can still turn up useful information but is losing depth. Full Stack Recruiter says the change did not appear to apply to every profile at first and was expected to roll out gradually. Pin, writing later, dates the change to January 2024.

The reason is contested. The Full Stack Recruiter newsletter says the official explanation is member privacy, while many in the industry believe the aim is to limit third-party recruiting tools.

X-ray still works better on GitHub. GitHub's Octoverse 2025 report says 180 million-plus developers now work and build on the platform. That is GitHub's own count. Pin says fewer than 13% of software engineering profiles in its own index have a public GitHub profile, so GitHub X-ray reaches a minority of engineers.

For related budget questions, see LinkedIn Recruiter vs Indeed vs ZipRecruiter and which job boards bring real hires.

How do platforms differ in Boolean support?

Support varies by platform, and a string that works in one place can fail silently in another. The table combines Google's and SeekOut's own documentation with Pin's reading of LinkedIn's documentation. Pin's claims are single-source and vendor-written, so check them against LinkedIn's help page before relying on them.

The wildcard row contains a real source conflict. Workable says an asterisk shortens a LinkedIn job-title search, for example administr*. Pin says LinkedIn ignores the asterisk entirely and tells readers to strip wildcards before pasting a Google string into LinkedIn. Test it on your own account before you trust either claim.

Pin also reports that Indeed handles AND, OR and quotes reliably, but that NOT and parentheses can give inconsistent results. If you run searches inside an ATS, check its own help page. Our ATS comparison for small businesses covers tools where this matters.

What Boolean mistakes waste the most time?

The most common errors are lowercase operators, missing quotes, missing parentheses and over-broad exclusions. Pin lists these in its cheat sheet, and the first three match the basics in the SeekOut and Workable guides.

• Lowercase operators. SeekOut says its tool needs capitals. Pin says LinkedIn treats lowercase "and" as a normal word.
• Unquoted phrases. machine learning engineer without quotes matches the three words anywhere on a profile.
• Missing parentheses. "data scientist" OR "data analyst" AND Python can be read as any data scientist, regardless of Python, per Pin.
• Stacked NOT terms. Pin warns that excluding "manager" also removes candidates who managed a team. Scan what you excluded.
• One title only. Workable's example: a narrow title misses valid profiles.
• Trusting the first page. Workable warns about false positives, such as someone who studied in Boston but does not live there.

One compliance point: these techniques are for manual searching. Check each platform's terms before you automate any search or collect profile data at scale.

Is Boolean still worth learning when AI search exists?

Yes, as a precision tool, though the evidence that AI search beats it is thinner than vendor guides suggest. SeekOut, which sells AI search, says Boolean remains useful for precision queries but that AI has surpassed it for most uses. That is a vendor claim.

One public study is an arXiv preprint, Evaluating AI Recruitment Sourcing Tools by Human Preference, submitted on 3 April 2025. The authors compared AI tools with LinkedIn Recruiter and their own system, Pearch.ai, across 48 queries and 1,735 results. They report that AI tools outperformed LinkedIn Recruiter on relevance, with Pearch.ai highest. The queries were written in plain language, such as a backend engineer in San Francisco with five years at startups, so the paper does not test hand-built Boolean strings.

What's still unclear?

• What top headhunters actually use. No public source documents this, so the strings above are vendor and trainer templates.
• LinkedIn wildcards. Workable and Pin disagree, so check LinkedIn's own help page.
• How much X-ray still works on LinkedIn. The 2024 reports describe a gradual change. The current extent has not been measured in public.
• Whether AI beats Boolean. The preprint compared AI tools with LinkedIn Recruiter, and its authors sell the top-ranked tool. It is a version 1 preprint.
• Template quality. Role strings from vendors are examples. None has been validated against hiring outcomes in a public source.