Overview of Gem’s AI features

AI paid add-ons

These advanced AI products are available for purchase through your Gem Account Manager:

• AI Sourcing: Searches across 800M+ public profiles and your organization's existing Gem CRM/ATS data to identify top prospects based on user-input scoring criteria. Here’s a 5-minute demo of AI Sourcing in action.
• Provides AI-powered search assistance, such as suggesting criteria based on your job description, generating company descriptions when you hover over a candidate's profile, or interpreting a user input like "publicly traded fintech companies" to suggest relevant organizations to source from.
• Candidates are returned with transparent percentage match scores, and any recruiter feedback is not used for AI training.
• Important safeguards include: PII is stripped from profiles before AI processing, warnings appear for subjective criteria that may introduce bias, and all scoring criteria must be approved by recruiters before use.
• AI-powered Application Review: Prioritizes active applicants in your ATS based on user-input scoring criteria. Here’s a 5-minute demo of AI-powered Application Review in action.
• This is an optional setting that can be turned on when reviewing applications for individual jobs.
• AI scoring is based entirely on recruiter-defined criteria, with the same safeguards as AI Sourcing: PII removal before processing, bias warnings for subjective inputs, and transparent percentage match scoring.
• The system helps prioritize applicant lists but does not make any advancement or rejection decisions — all hiring actions remain with recruiters.
• Note: Applications without attached resumes are not scored by AI and will appear in your results without a match percentage. This ensures no candidates are hidden from view.

Complimentary AI features

These features are available to all Gem customers:

• Sequence Template Generation: Generates customizable email templates and personalization tokens to assist recruiters with outreach messaging, while maintaining full recruiter control over final message content.
• AI Custom Token in Sequences: Helps recruiters further personalize their outreach and save time by automatically suggesting a relevant sentence based on the prospect/candidate profile.
• AI Skill Suggestion for Application Review: Suggests keywords to search on applications based on the ATS job description.
• AI Prospect Search: Helps craft search queries based on your job description to find existing talent in your Gem CRM and integrated ATS.
• AI Company Descriptions: Automatically generates company overviews (including business focus, size, and industry) that appear when users hover over company names in candidate profiles, helping recruiters quickly understand a candidate's background without leaving the platform.
• AI Scorecard Summaries: Automatically summarizes detailed interview notes across interviewers for a specific Job.

What languages do these AI features support?

We are currently testing and developing in English only.

How can I provide feedback to Gem on these AI features?

Please reach out to your Account team with any feedback or questions.

Frequently asked questions

What powers Gem AI?

Gem’s AI features are powered by Microsoft Azure OpenAI, a trusted third-party vendor known for its enterprise-grade generative AI, security, and compliance. Read more about Microsoft’s Data, privacy, and security for Azure OpenAI Service.

How does Gem's AI match score algorithm work?

Gem's match score algorithm uses recruiter-defined criteria to evaluate candidates through our Azure OpenAI LLM. Here's how it works:

• Recruiters input specific job-related criteria
• The LLM scores each criterion individually against the candidate's resume or public profile
• The final match score displayed is an average across all individual criteria scores
• All scoring is transparent, with percentage matches and explanations for each criterion
• Recruiters maintain full control to adjust criteria and re-score at any time

Is customer data used to train your AI?

No. Gem does not use customer data to train the AI. Instead, we use pre-trained models with Microsoft Azure OpenAI.

• GDPR Compliant: The standard data retention period extends to 90 days after the termination of the last active contract between Gem and a customer, but the customer can set custom retention periods in the Gem admin interface.
• PII Redaction: We redact personally identifiable information (PII), such as names, locations, and LinkedIn handles, before sharing it with our AI.
• Vendor Privacy: Our vendor Microsoft OpenAI does not store or use Gem’s customer data for training. OpenAI may securely retain API inputs and outputs only for up to 30 days to detect abuse and maintain service quality. Read more about Microsoft’s Data, privacy, and security for Azure OpenAI Service.
• Data: All data processed by Gem’s AI features is limited to your Gem account. At no time will other Gem customers have access to your AI inputs or outputs.
• Data Requests: Candidates can submit data subject access requests to the Gem Support team at support@gem.com.

How does Gem prevent and detect bias in AI recommendations?

Gem proactively addresses bias through multiple approaches:

• Proactive guardrails: We remove personally identifiable information (PII) like names, locations, and demographic indicators before AI processing, and provide warnings for subjective criteria that may introduce bias.
• Third-party bias audits: We've conducted a successful bias audit with BABL and will undertake annual audits to continuously monitor for potential bias in our AI recommendations.
• Human oversight: Gem AI assists with decision-making but does not make hiring decisions. Recruiters maintain full control over criteria input and can edit AI suggestions at any time.
• Transparent scoring: Our AI provides numeric scores with explanations, allowing users to understand and validate how candidates were evaluated.

If you have concerns about AI recommendations, please contact your Gem Account Manager or our Support team at support@gem.com.

Is Gem SOC2 Type 2 compliant?

Gem’s infrastructure, including AI features, undergoes annual audits. SOC 2 audit reports are available under NDA. Contact your Gem Account Manager for access.

How does Gem comply with California's FEHA regulations on automated decision systems?

California's October 2025 FEHA amendments clarify that anti-discrimination laws apply to all employment decisions, including those using automated decision systems (ADS).

While Gem's AI features are designed as decision-support tools rather than autonomous decision-makers, we've structured our practices to support customer compliance with FEHA requirements, whether or not our systems are ultimately classified as automated decision systems under California Code of Regulations Title 2, Section 11008.1:

Our compliance approach includes:

• Human control: Recruiters define all scoring criteria, approve inputs before use, and maintain exclusive authority over hiring decisions. No candidate is automatically advanced or rejected by AI.
• Anti-bias measures: We strip PII (names, locations, demographic indicators) from profiles before AI processing and provide real-time warnings for subjective criteria that may introduce bias.
• Third-party auditing: We've completed a comprehensive bias audit with BABL specifically for our AI-powered Application Review feature and commit to annual audits going forward, demonstrating the ongoing anti-bias testing that Section 11009(f) recognizes as relevant evidence in discrimination claims.
• Data control: Customers can configure retention periods through the Gem admin interface to meet the four-year recordkeeping requirements for automated decision system data under Section 11013(c).
• Job-relatedness: All AI scoring is based entirely on recruiter-defined criteria tied to specific job requirements, aligning with FEHA's business necessity and job-relatedness standards under Sections 11017 and 11010(b).

Whether classified as an ADS under California regulations or not, Gem's systems are built on principles of fairness, transparency, and human oversight to support compliant, non-discriminatory hiring practices.

How does Gem address NY Local Law 144?

NY Local Law 144 requires employers and employment agencies that utilize an Automated Employment Decision Tool (AEDT) to conduct a bias audit on such tools to ensure they promote fairness and transparency in the hiring process.

Gem’s AI solutions are designed to accelerate decision-making, not make decisions, and therefore we do not consider that our tools will fit under the definition of an AEDT under NY Local Law 144. Nonetheless, for our AI-powered App Review solution specifically, we proactively conducted a bias audit to ensure all customers can confidently incorporate AI into their hiring processes, regardless of location or regulatory requirements, and in accordance with their own interpretation of NY Local Law 144.

We’ve conducted a successful bias audit with third-party partner BABL and will undertake an annual audit as part of our ongoing compliance process. In addition, we have robust internal governance in place to ensure our platform maintains proper security and procedural standards.

Has Gem been impacted by Workday’s recent lawsuit?

No. We're actively monitoring the Mobley v. Workday litigation. This case is still in early stages, and the specific circumstances may differ significantly from Gem's approach, which emphasizes human control and decision support rather than automated decision making. We continue to apply rigorous standards, including proactive bias auditing, PII redaction where appropriate, and transparent scoring with human oversight. To date, we haven't had customers cease using AI-based products in response to this litigation, and we remain committed to partnering with customers to ensure continued compliance as laws evolve.

What about other regulations?

Gem proactively monitors AI laws and regulations, including, the EU AI Act, Colorado’s AI Act, and other proposed legislation throughout the US, and will continue to partner with customers to support them in remaining compliant with applicable laws as they come into effect, such as:

• EU AI Act
• California FEHA amendments (automated decision systems)
• CCPA draft regulations
• Colorado AI Act

How does Gem develop and govern AI features?

Rigorous Testing

All Gem features undergo manual and automated testing in our development environments as part of our SDLC and change management processes. This includes any AI-powered features.

Security & Governance

Our internal AI governance committee is responsible for reviewing all AI features and ensuring they are built in accordance with the relevant guidelines. Our AI features have been developed in accordance with our existing audited security controls, which include a complete security review of our subprocessors.

Gem’s commitment to responsible AI

At Gem, we believe artificial intelligence should empower people—not replace them. Our AI solutions are designed to assist talent teams in making faster, more informed decisions while ensuring fairness, transparency, and control at every step. Our approach to responsible AI is guided by the following principles:

Transparency and control

Gem AI does not make hiring decisions. We let recruiters control what is inputted into the AI (such as criteria to score a profile) and display their input with objective numeric scores and explanations that provide context on how a candidate matched. Users can edit what they input into the AI at any time.

Proactive guardrails

We’ve added proactive guardrails to ensure responsible AI use. This means removing personally identifiable information (PII) from profile data before sending it to the AI for scoring, and we provide warnings for subjective inputs, like “demonstrates leadership skills,” that may be challenging for AI to interpret consistently.

Enterprise-grade compliance

Our platform and compliance framework are reviewed by internal governance and third-party audits to ensure proper security and procedures. We’ve conducted a successful bias audit with our partner, BABL, and will undertake an annual audit as part of our process.

Next-gen technology

Unlike legacy AI vendors that use outdated machine-learning models, Gem uses the latest generative AI, delivering superior performance. This next-gen technology delivers better accuracy, more transparency & control, and faster calibration.

Have any issues or questions on this topic? Please feel free to contact your dedicated Gem Customer Success Manager directly or our Support team at support@gem.com.

• Overview of Gem’s AI features
• AI paid add-ons
• Complimentary AI features
• What languages do these AI features support?
• How can I provide feedback to Gem on these AI features?
• Frequently asked questions
• What powers Gem AI?
• How does Gem's AI match score algorithm work?
• Is customer data used to train your AI?
• How does Gem prevent and detect bias in AI recommendations?
• Is Gem SOC2 Type 2 compliant?
• How does Gem comply with California's FEHA regulations on automated decision systems?
• How does Gem address NY Local Law 144?
• Has Gem been impacted by Workday’s recent lawsuit?
• What about other regulations?
• How does Gem develop and govern AI features?
• Rigorous Testing
• Security & Governance
• Gem’s commitment to responsible AI
• Transparency and control
• Proactive guardrails
• Enterprise-grade compliance
• Next-gen technology