The short answer
Bullhorn fake candidates and spam applications mostly enter through the same door: web responses from your job board, LinkedIn and Indeed. Staffing desks can cut most of the junk with Bullhorn's own tools, used in order. Triage web responses before converting them to candidates. Let Record Duplicate Checking catch repeat records. Filter and review by source. Turn on Amplify Verify (Early Access) for identity and behavior signals, and treat each result as a signal, not a verdict. On Amplify Edition, add Amplify Screen with an Automation knockout rule, plus a new-applicant follow-up. Then use REST event subscriptions to send new candidates to a screening step. None of this verifies identity on its own.
Where the junk comes from
Staffing desks feel application volume more than most. LinkedIn says 11,000 of its members apply for jobs every minute. In LinkedIn's January 2025 research, 73% of HR professionals said less than half of the applications they get meet all the criteria. Greenhouse's November 2025 survey found 34% of recruiters spend up to half their week filtering spam and junk applications.
Then there is outright fraud. The FBI warned in January 2025 that North Korean IT workers had used AI and face-swapping in video interviews, and that the same phone numbers, often VoIP numbers, appeared across different resumes. The FBI's 2022 warning named remote IT, programming, database and software jobs, the same roles IT staffing desks fill every day. For the interview side of that problem, read our guide to detecting proxy candidates in remote interviews.
In Bullhorn, applications typically arrive three ways. Applications from your career site, including Bullhorn's free, self-hosted Open Source Career Portal, land as web responses. LinkedIn Apply Connect sends LinkedIn applications in with Source set to LinkedIn. Indeed Syndication and Indeed Apply are included automatically, with Source set to Indeed. One note on the Open Source Career Portal: Bullhorn's overview page doesn't mention spam or CAPTCHA protection, so if you self-host it, plan your own bot protection.
Bullhorn's tools at a glance
Triage web responses before they become candidates
Every application in Bullhorn starts as a web response with a parsed resume. That is your first and cheapest checkpoint, because nothing has polluted your database yet. For each response you can Convert it into a candidate, Reject it, or Delete it. Rejected responses are kept for reporting, so Reject is usually the better choice for junk: you keep a record of what came in and from where, which helps when you tune your filters later.
If a response belongs to someone already in your database, use Parse as Existing to attach it to that record instead of creating a new one. That one habit prevents a lot of duplicate cleanup later.
Let duplicate checking do its job
Record Duplicate Checking runs when a candidate, contact, lead or company is added. It compares last name plus the first two letters of the first name plus Email 1. A match on Email 1 alone shows a warning. The check includes archived records, which matters for staffing: a candidate you placed or rejected two years ago still counts.
The gap is a new email address. Someone mass-applying with fresh emails, or a fake identity with a new inbox, won't match. Phone numbers help here. The FBI's note about reused VoIP numbers suggests a simple habit: search Bullhorn for the phone number on a suspicious record and see whether it appears on other candidates.
Filter LinkedIn and Indeed applications by source
LinkedIn Apply Connect matches an incoming applicant to an existing candidate by LinkedIn ID (URN), or by full name plus email. Indeed applications arrive through Indeed Syndication and Indeed Apply. Both set a Source value, so you can filter your web responses to Source = LinkedIn or Source = Indeed and review each channel as its own queue.
Reviewing by source shows you which channel brings the most junk for each job type, and it matters for the next step, because Amplify Verify returns No Signal Data for these two sources.
Catch Bullhorn fake candidates with Amplify Verify
Amplify Verify looks for three kinds of signals:
- Identity signals, such as VPN or proxy use, device or network reuse, and disposable email addresses
- Behavioral signals, which need a JavaScript snippet on your apply page
- Misrepresentation signals, such as copy-paste and AI-generated content
Each applicant gets a result of Clear, Flagged, Needs Review or No Signal Data. Applications from LinkedIn and Indeed return No Signal Data. Bullhorn's own wording is the right way to use it: a signal, not a verdict. A Flagged result means a recruiter should look closer, maybe with a short call, not that the person is fake. Bullhorn lists Amplify Verify as currently in Early Access, so check availability with your Bullhorn contact.
Screen at scale with Amplify Screen and knockouts
Amplify Screen runs AI chat and voice screening. Bullhorn generates the questions when the job is added, and each screen returns a score, a summary, strengths and gaps, and a transcript. It needs Amplify Edition.
With Amplify Screen in place, Bullhorn Automation can update a submission's status based on screener answers. If a candidate gives a knockout answer, the automation sets the submission to Rejected. It needs an Always Updated submission list. Keep knockouts for real requirements, like a license the client requires or availability for a set shift, so you don't lose good candidates to a badly worded question.
Bullhorn Automation also has a New Applicant Follow-up use case. When a candidate applies, it can send an autoresponder, notify the recruiter, update status or score, and create tasks. Fast, consistent follow-up keeps real candidates warm while the junk gets sorted.
Route new candidates to Carolina with event subscriptions
Bullhorn's REST API offers event subscriptions for INSERTED, UPDATED and DELETED events. Your system polls for events, up to 100 per request. Each database can hold 15 subscriptions, and unread events expire after 7 days, so poll often. A subscription to new candidate records is a clean trigger for a screening step that runs after your own checks.
Carolina is NTRVSTA's AI recruiting agent, and she connects to Bullhorn through our Bullhorn integration. She scores every applicant against your scorecard and writes down her reasons for each candidate, so a recruiter can see why someone ranked where they did. She runs live first-round interviews by video or voice at any hour, and the questions adapt to each answer with follow-ups. Scorecards and transcripts are written to Bullhorn. You set each action to suggest, ask first or run on its own, and offers always wait for a person. More on how staffing teams set this up is on our staffing page.
Be clear about what this does and doesn't do. Live follow-up questions make memorized or relayed answers harder to sustain and easier for a human reviewer to notice, but that is not identity verification. Keep Amplify Verify and your duplicate checks in place, and add an ID check from your background-check provider before a placement starts. NTRVSTA connects to Checkr, Sterling and HireRight. For the broader approach, see how to stop AI resume spam, and size the problem with the application spam cost calculator.
Where NTRVSTA isn't the right fit: if your desks are on Amplify Edition and Amplify Screen plus knockout automation already handles your screening volume, you may not need another tool. For very high-volume hourly staffing where SMS apply and chat scheduling are the core flow, conversational hiring tools built for that flow fit better. NTRVSTA is also newer and smaller than the large incumbents. On Ashby, Greenhouse or Lever? See our Ashby, Greenhouse and Lever guides.
Questions
Does Amplify Verify reject fake candidates in Bullhorn?
No. Bullhorn describes an Amplify Verify result as a signal, not a verdict. It labels applicants Clear, Flagged, Needs Review or No Signal Data, and a recruiter decides what to do. Bullhorn lists the feature as currently in Early Access.
Why do LinkedIn and Indeed applicants show No Signal Data in Amplify Verify?
Bullhorn's docs say applications that arrive through LinkedIn and Indeed return No Signal Data. Behavioral signals also need a JavaScript snippet on your own apply page. Review those sources with duplicate checks and screening instead.
What does Bullhorn duplicate checking match on?
When a candidate, contact, lead or company is added, Bullhorn compares last name plus the first two letters of the first name plus Email 1. A match on Email 1 alone shows a warning. Archived records are included in the check.
Do we need Amplify Edition for knockout automation?
Yes. Bullhorn's knockout recipe sets a submission to Rejected based on screener answers, and it needs Amplify Screen, which requires Amplify Edition. It also needs an Always Updated submission list in Automation.
Running a Bullhorn desk and want to see Carolina score your applicants against a client's scorecard? Book a demo with your current workflow in hand.
Sources
- Bullhorn: managing web responses
- Bullhorn: understanding record duplicate checking
- Bullhorn: reviewing LinkedIn web responses
- Bullhorn: Indeed Syndication
- Bullhorn: Open Source Career Portal overview
- Bullhorn: using Amplify Verify
- Bullhorn: Amplify Screen quick start
- Bullhorn Automation: update submission based on screener questions
- Bullhorn Automation: new applicant follow-up
- Bullhorn: REST API event subscriptions
- Bullhorn: REST API documentation
- LinkedIn: hiring product update (2025)
- LinkedIn: job seeker and HR research (January 2025)
- Greenhouse: hiring manager and job seeker survey (November 2025)
- FBI IC3: North Korean IT workers (January 2025)
- FBI IC3: deepfakes and stolen PII used to apply for remote work positions (June 2022)




