Using Scoring Guidance with Amplify Verify

Amplify Verify is currently in Early Access. For more information, speak to your Bullhorn representative.

Amplify Verify uses a default set of instructions to evaluate candidate signals and return a risk assessment. The Scoring Guidance setting lets you add your own instructions on top of the default — for example, to tell Amplify Verify which signals to treat as more or less significant for your organization.

Scoring Guidance is applied across your whole organization and can only be configured by an Admin user. Your instructions work alongside the default: you are adding to it, not replacing it.

What the Default Does

Out of the box, Amplify Verify is configured to act like a fair, experienced trust analyst, weighing signals in context rather than applying a rigid checklist. The default instructs it to:

  • Evaluate candidates across four signal categories, each returning its own verdict:

    • Identity: VPN and proxy use, virtual machine detection, anti-fingerprint browser detection, browser identity tampering, devices or IP addresses shared with other candidates, disposable email addresses, and IPs that appear on threat or spam blocklists.

    • Behavioral: Bot detection scores, honeypot field interactions, human interaction events (scrolling, pointer movement, clicks, and text editing), suspiciously fast field completion, forms submitted immediately on opening, application rate (number of applications in a 24-hour window), and session burst patterns. Recruiter notes indicating genuine engagement, such as a completed call or interview, are also factored in as a positive signal.

    • Misrepresentation: Detection of a hidden AI-instruction keyphrase in submitted answers (whether typed or pasted), text pasted after leaving the application page, and other signs that answers were not written by the candidate in the moment.

    • Consistency: Conflicts or unexplained changes between resume versions, and whether the candidate's sessions originated from countries that match their profile location.

  • Judge signals by prevalence, not just presence. A signal appearing once across many sessions is treated as weak evidence; the same signal appearing repeatedly is treated as meaningful.

  • Default to Clear. Amplify Verify only flags a candidate when there is concrete, material evidence with no obvious innocent explanation. Common behaviors such as VPN use, copy-pasting, and virtual machines on corporate devices are not treated as suspicious on their own.

  • Protect candidate anonymity. Amplify Verify does not have access to the candidate's name, email address, phone number, or any other personally identifying information. It evaluates signals only.

  • Return a structured, plain-language result that includes a verdict and supporting bullet points for each category, plus a short overall summary for the recruiter.

Adding Scoring Guidance

To add or update your Scoring Guidance:

  1. Go to Menu > Admin > Amplify Admin.

  2. Click the Labs tab.

  3. Click the settings icon () to the right of Amplify Verify.

  4. Enter your instructions into the Scoring Guidance field.

  5. Click Save.

Scoring Guidance Examples

The following examples show how you might use Scoring Guidance to better reflect your organization's hiring context.

Disregard VPN Use

Default behavior: VPN use is noted as an informational signal and may contribute to an Identity flag when combined with other signals.

When to change this: If your candidates commonly connect via VPN, for example because they are applying from corporate networks or are privacy-conscious professionals, you may want Amplify Verify to disregard VPN detection entirely.

Example instruction:

"Do not treat VPN or proxy detection as a signal of concern, regardless of confidence level or how many sessions it appears in."

Disregard AI-Assisted Writing Signals

Default behavior: Amplify Verify looks for signs that answers were AI-generated or not written by the candidate in the moment, including detection of a hidden AI-instruction keyphrase and text pasted after leaving the page.

When to change this: If you hire for roles where AI-assisted writing is acceptable, or if your application process explicitly permits candidates to use AI tools, you may want to remove Misrepresentation signals from consideration.

Example instruction:

"Do not flag candidates for AI-assisted content or pasted answers. Disregard all Misrepresentation signals and return that category as Clear."

Focus Only on Automated Submission Signals

Default behavior: Amplify Verify evaluates all four signal categories and returns a holistic verdict.

When to change this: Some organizations may only want to screen for bot activity and automated submissions, and prefer not to surface signals about identity, content, or consistency.

Example instruction:

"Focus your assessment on the Behavioral category only. Return Identity, Misrepresentation, and Consistency as Clear regardless of the signals present, and base the overall result on Behavioral signals alone."