AI prompt for boolean search engine

Build Boolean search strings for a recruiting or research search. Treat the search as a recall-and-precision problem: start broad enough to find qualified people with nontraditional titles, then provide controlled narrowing options. Follow applicable company policy and do not use protected characteristics, proxies for them, or discriminatory exclusions.

Search brief:
- Target role and level: [TARGET ROLE]
- Must-have capabilities: [MUST HAVE CAPABILITIES]
- Helpful but optional capabilities: [PREFERRED CAPABILITIES]
- Titles, adjacent titles, and terminology: [TITLE AND TERM NOTES]
- Geography, work authorization, or other policy-approved constraints: [APPROVED CONSTRAINTS]
- Search platform and character limits: [SEARCH PLATFORM]

Produce four items:
1. A core Boolean string using title and skill concept blocks.
2. A broader discovery string designed to surface adjacent backgrounds.
3. A narrower string that adds only job-related preferred criteria.
4. A readable logic map showing each synonym group and why it is included.

Use parentheses correctly, quote multiword phrases only when that improves precision, and avoid arbitrary exclusions such as “junior,” “assistant,” or employer names unless I explicitly authorize them and they are job-related. If the platform has different Boolean behavior, adapt syntax and state the assumption. Include a testing plan: review the first 25 results, identify false positives and false negatives, change one concept block at a time, and log the revision.

Provide a short fairness and policy note identifying any input that could create an inappropriate restriction. Do not claim a candidate is qualified based on keywords alone. Ask up to 3 clarifying questions only if a required input is missing.

Before answering, validate balanced parentheses and ensure every exclusion is supported by the stated job requirements or approved constraints.

Create a Google X-ray Boolean search for candidates on [TARGET SITE]. Role: [ROLE]. Required skills: [REQUIRED SKILLS]. Title variants: [TITLE VARIANTS]. Approved location constraints: [APPROVED LOCATION RULES]. Return three copyable queries using site: syntax, quoted phrases only where appropriate, and grouped synonym blocks. Explain which terms may reduce recall. Do not use protected traits, personal contact data, or demographic proxies. Add a first-page review checklist for relevance and false negatives. Ask up to 3 questions only if an input is missing. Validate parentheses and confirm every constraint is job-related before answering.

Create Boolean searches to map the talent market for [ROLE FAMILY] in [APPROVED REGION]. Core capabilities: [CAPABILITIES]. Known title variants: [TERMS]. Platform: [PLATFORM]. Produce a broad search, an industry-focused search, and a skills-focused search, plus a table explaining each concept block. Flag title terms likely to be ambiguous and suggest a manual sample-review process. Do not infer candidate quality from profiles or introduce demographic filters. Ask up to 3 clarifying questions only if required inputs are missing. Check syntax and identify any constraint not directly supported by the brief.

Diagnose and improve this Boolean search: [CURRENT STRING]. Target role: [ROLE REQUIREMENTS]. Examples of relevant results: [RELEVANT EXAMPLES]. Examples of irrelevant results: [IRRELEVANT EXAMPLES]. Platform: [PLATFORM]. Explain likely recall and precision problems, then provide a revised core string and one alternate test string. Change one logic block at a time and label each change. Preserve job-related inclusion criteria; do not add protected-characteristic proxies or unsupported exclusions. Ask up to 3 questions only if necessary. Before answering, check balanced parentheses and explain any platform-specific syntax assumption.