Gem’s AI Sourcing works best when your search is scoped correctly: broad enough to capture strong candidates, but focused enough to prioritize top matches.
Use this guide to troubleshoot common issues.
If you have too few profiles
Common Causes may include:
- Too many restrictive filters (degree, company size, years of experience)
- Only one job title listed (too specific)
- Boolean search too narrow
Best practices when you too few profiles:
- Move requirements (e.g., degree, company size, years of experience) out of Filters and into Criteria instead to:
- Allow for ranking the profiles which best fit your ideal prospect persona, without overlooking prospects who would fit your role.
- Allow Gem AI to analyze the companies, degrees, or what you would’ve filtered by and instead have AI rank profiles by the types of companies, degrees, and other criteria which fit what you’re looking for. - Add multiple job title variants (e.g., “iOS Developer,” “Mobile Engineer,” “iOS Software Engineer”).
How to improve your search results for common scenarios:
| Bad Search | Improved Search | Why It's Better |
| Filter: Field of Study = Accounting |
Criteria: BA degree in Accounting, Finance, or related fields | Expands eligible talent pool and lets AI prioritize best matches |
| Boolean: "Tax Accounting" |
Criteria: At least 5 years of experience in accounting, tax audit, tax reporting, or corporate finance | Focuses on real skills not exact phrasing |
| Filter: Job Title = “Senior iOS Engineer” |
Titles Added: “iOS Developer,” “Mobile Engineer,” “Senior Mobile Engineer” |
Includes similar real-world titles used by candidates |
| Boolean: "backend development" |
Criteria: 4+ years of work experience in backend development, including building APIs or server-side systems, as shown in past job responsibilities |
Prevents overlooking talent who are a fit by having AI act as your assistant Recruiter to look for talent with the necessary work experience |
If you have too many profiles
Common Causes may include:
-
Only basic filters applied (e.g., job title = “Engineer”)
-
No must-have skills expressed
-
No Location specified
-
No years of experience specified
Best practices when you too few profiles:
-
Add must-have hard skills into the Skills field (e.g., “Python,” “AWS,” “Distributed Systems”).
-
Add Months/Years of Experience
- Months at current company (Experience & Tenure Filter)
- Experience & Tenure Filter (Experience & Tenure Filter)
- Experience for specific skills or in a particular industry (Criteria - Natural Language Search) -
Exclude noisy titles (e.g., “Intern,” “Assistant,” “Support Engineer”).
-
Include filters to find the best and high achieving talent, not just any talent with the skills.
How to improve your search results for common scenarios:
| Bad Search | Improved Search | Why It's Better |
| Filter: Field of Study = Accounting |
Criteria: BA degree in Accounting, Finance, or related fields | Expands eligible talent pool and lets AI prioritize best matches |
| Boolean: "Tax Accounting" |
Criteria: At least 5 years of experience in accounting, tax audit, tax reporting, or corporate finance | Focuses on real skills not exact phrasing |
Key reminders
-
Filters will block candidates completely.
- Criteria help the Gem AI rank candidates, leading to better matches
For more Gem AI best practices, see our AI Scoring Criteria Best Practices article.