> ## Documentation Index > Fetch the complete documentation index at: https://docs.alex.com/llms.txt > Use this file to discover all available pages before exploring further. # Cheat Detection Our Verify features are designed to give you confidence in your hiring process by providing insights into candidate integrity and identity. This guide will walk you through setting up, configuring, and interpreting the results of our Cheat Detection and Candidate Verification tools. *** ### Getting Started with Verify Get up and running with our integrity features to ensure a fair and secure interview process. #### Overview of Cheat Detection and Fraud Prevention Verify is a suite of tools that helps maintain the integrity of your interviews. It operates in two main ways: * **Cheat Detection:** Analyzes an interview's video, audio, and transcript to identify behaviors that may indicate cheating. This includes looking for external assistance, use of AI-generated answers, and suspicious candidate tendencies. * **Fraud Prevention (Candidate Verification):** Verifies a candidate's identity by checking the information they provide, such as their email, phone number, and professional profiles like LinkedIn. This helps you confirm that candidates are who they say they are. #### Enable Verify Features in Company Settings Both Cheat Detection and Fraud Detection can be enabled or disabled for your entire company within your company's settings. *** ### Configuring Cheat Detection You can adjust how Cheat Detection works to better fit your company's needs. #### Set Cheat Detection Sensitivity Thresholds You can define a "Cheat Detection Threshold" in your interview settings. This is a score from 0 to 100. If a candidate's cheating likelihood score exceeds this number, the interview will be flagged for your review. #### Understand Monitoring Capabilities Our system analyzes multiple aspects of an interview to detect potential cheating: * **Video:** For video interviews, the system monitors for the presence of multiple people and analyzes the candidate's head and eye movements. * **Audio:** The audio is analyzed to detect other voices that might be assisting the candidate and to check for the use of voice-changing technology. * **Transcript:** The interview transcript is analyzed to identify if the candidate's responses appear to be generated by AI tools. For phone interviews, only the audio and transcript are analyzed. *** ### Interpreting Cheat Detection Analysis The cheat detection analysis is broken down into several components to give you a comprehensive view. The results are categorized into three probability levels based on a score from 0-100. #### Understand Cheating Probability Levels * **Unlikely (0-40):** No significant evidence of cheating was found. * **Possible (41-75):** Some suspicious behavior was detected, but the evidence is not conclusive. * **Likely (76-100):** Clear evidence of cheating was detected. #### Analyze Eye and Head Tracking Tendencies This analysis, labeled **Candidate Tendencies**, monitors a candidate's eye and head movements. It looks for patterns that might suggest the candidate is reading answers from an off-screen source rather than answering from memory. #### Detect External Assistance This analysis looks for help from outside sources and is presented in two parts: * **External Assistance:** This part analyzes the video feed to detect if other people are present and potentially assisting the candidate. * **Voice Analysis:** This part listens to the interview audio for signs of other voices providing answers or the use of voice-changing software. #### Identify AI-Generated Responses from Transcript Analysis The **Transcript Analysis** reviews the candidate's spoken answers. It identifies responses that have characteristics of being generated by AI models, such as being overly structured or generic, to distinguish them from genuine, personal experiences. *** ### Candidate Verification (Fraud Detection) Fraud Detection helps you verify a candidate's identity and assess the risk of fraudulent activity. #### Interpret Fraud Risk Scores The overall fraud risk is determined by a score and categorized into three levels: * **Low (80-100):** The candidate's information appears to be legitimate. * **Medium (60-79):** Some inconsistencies were found that may warrant a closer look. * **High (0-59):** Significant red flags were detected, indicating a high risk of fraud. #### Perform Identity Checks The system performs several checks on the candidate's provided information: * **Email Verification:** Checks if the email is from a disposable service, its deliverability status, and its estimated age. * **Phone Verification:** Determines if the phone number is a VOIP (Voice over IP) number, and checks the Caller ID Name (CNAM) to verify it matches the candidate's name. * **IP Verification:** Checks if the candidate is using a VPN or proxy and identifies their general location (city and region). #### Verify Professional Identity To help confirm a candidate's professional background, the system performs the following: * **LinkedIn Verification:** Analyzes the candidate's LinkedIn profile, looking at its age and the number of connections. * **Resume Matching:** Compares the candidate's resume against their LinkedIn profile to check for consistency in their work history and experience. *** ### Managing Verify Results in Reports You have control over how Cheat Detection and Candidate Verification data appears in your reports. #### Configure PDF Report Layouts for Verification Data In your settings, you can customize the layout of your PDF reports. You can choose to include or exclude the **Cheat Detection** and **Candidate Verification** sections. You can also drag and drop these sections to change their order in the final report. #### Control Verify Visibility in Shared Reports Similar to PDF reports, you can also control the visibility of the **Cheat Detection** and **Fraud Detection** sections in the public, shareable reports you send to others. This can be configured in your report layout settings.