The short answer
Ashby spam applications come in three kinds: bot submissions, people mass-applying with AI-written resumes, and fake or proxy candidates. Ashby has a native tool for each layer, and the order you turn them on matters. Start with job board spam protection to block bots. Add Application Limits to stop repeat applying. Then set auto-reject rules and AI-Assisted Application Review criteria to clear unqualified volume. Turn on Ashby fraud detection with a review stage so a person checks risky candidates. Finally, use the applicationSubmit webhook to route clean applications to a screening step. Several of these depend on your plan, and none of them verifies identity on its own.
Why Ashby teams are seeing more of this
Volume is the first problem. LinkedIn says 11,000 members apply for jobs every minute, and 37% of job seekers in its January 2025 research said they were applying to more jobs than ever. Greenhouse put the average at 222 applications per job opening in Q1 2024, almost three times the level at the end of 2021. Greenhouse's November 2025 survey found 34% of recruiters spend up to half their week filtering spam and junk applications.
Fake candidates are the second problem. Gartner predicts that by 2028, one in four candidate profiles worldwide will be fake. The FBI warned in January 2025 that some remote job applicants have used AI and face-swapping in video interviews. If that risk is yours, read our guide to detecting proxy candidates in remote interviews alongside this one.
Ashby's spam and fraud tools at a glance
Step 1: Turn on job board spam protection
Job Board Application Spam Protection adds bot protection to your Ashby-hosted job board. An Org Admin picks one of four levels: Strict, Permissive (Recommended), Less Permissive, or No Protection. Start with the recommended setting. Move to Strict only if bot traffic keeps getting through, and watch your applicant numbers after any change, since a tighter setting can turn away real people too.
One related trap: unlisting a posting doesn't stop applications. Ashby's docs say unlisted postings can still be scraped and applied to. If a role is closed, unpublish it.
Step 2: Set Application Limits
Application Limits control how often one person can apply. You can also block rejected candidates from reapplying, either to the same job or to any job. Blocked attempts show up in a Blocked Application Submissions tab, so you can check what was stopped. Ashby's own help article on multiple applications from one candidate calls Application Limits "the best tool to address this."
Two caveats. Limits match people by email, so someone using a new address gets through. And they don't apply to applications that come through Indeed. Application Limits are on Legacy Plus, Plus and Enterprise plans, not Foundations.
For people who slip past with a second email, use duplicate detection. Ashby flags likely duplicates by email, name and social links such as LinkedIn and GitHub, and you choose Merge Profiles or Not a Duplicate. Merging is irreversible, so check before you click.
Step 3: Add auto-reject rules and AI-Assisted Application Review
Auto-Reject Applications rejects applicants based on their application form answers. Conditions include Is empty, Equals, Contains, Starts with, Ends with and Similar to. You can send an optional rejection email after a set number of business days. Rules are not retroactive, so they only apply to applications that arrive after you save them. Keep auto-reject for true requirements, such as work authorization or a required location. Enterprise customers can apply questions and rules across every matching posting with Global Application Questions and Global Auto-Reject Rules.
For judgment calls, use AI-Assisted Application Review. You write criteria per job. The AI marks each application Meets or Does not Meet, with its reasoning, and a person decides whether to advance or reject. It's available on all plans and uses AI credits. Ashby also says the analysis flags blank or unrelated answers to your application questions as not meeting criteria, which helps with copy-and-paste mass applicants.
Step 4: Turn on Ashby fraud detection with a review stage
Candidate Fraud Detection analyzes new applications as they are submitted, using device, IP, email and phone signals. Enable it in admin settings, then turn it on per job. Checks can run automatically, or a recruiter can run Check for Fraud by hand. Ashby says it's available across all plans and uses two AI credits per candidate.
Reviewing is manual by design. For each flagged candidate, choose:
- Mark Fraud, which opens the archive flow with the reason "Rejected by Organization: Fraudulent Candidate"
- Not Fraud, which adds the label Fraud Signals - Cleared
- Unsure, which adds the label Fraud Signals - Needs Review
Use the filter Multiple Fraud Signals - Unreviewed to build a daily queue. You can also add a dedicated security-review stage, as Ashby's own fraud tooling recipe does, so flagged candidates don't sit in the main pipeline. Give that stage one owner and a written checklist, and treat Unsure as "talk to the candidate," not "reject."
Step 5: Route clean applications with the applicationSubmit webhook
The applicationSubmit webhook fires when someone applies or is added manually. It does not fire for bulk imports or applications from integrations. For job board applications, the payload includes submitterClientIp and submitterUserAgent, and you can verify requests with the optional Ashby-Signature header.
Ashby documents a recipe for teams that want their own checks:
- Receive the applicationSubmit webhook.
- Call candidate.listClientInfo to get client details.
- Run them against your threat intelligence source.
- Write the result back with candidate.createNote, move risky applications with application.change_stage to your security-review stage, and tag them with candidate.addTag.
Applications that pass can move on to a screening stage. That is where a first-round interview fits, whether you use Ashby's own AI Interviewer for automated async screening (an opt-in feature under Admin, Organization Setup, Opt-In Features) or connect a recruiting agent.
Where Carolina fits after Ashby's filters
Carolina is NTRVSTA's AI recruiting agent, and she connects to Ashby through our Ashby integration. Once Ashby's filters have done their job, she scores every remaining applicant against your scorecard and writes down her reasons for each candidate. She can run live first-round interviews by video or voice at any hour, where the questions adapt to each answer with follow-ups. Scorecards and transcripts are written back to Ashby, so your team reviews everything in one place. You set each action to suggest, ask first or run on its own, and offers always wait for a person.
Be clear about the limit. 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 Ashby's fraud review in place, and add an ID check from your background-check provider. NTRVSTA connects to Checkr, Sterling and HireRight. For the wider playbook, see how to stop AI resume spam, or estimate what the junk costs you with the application spam cost calculator.
Where NTRVSTA isn't the right fit: if spam protection, Application Limits, auto-reject and Ashby's fraud review already keep your queue manageable, you may not need another tool. The same goes if Ashby's AI Interviewer covers the async screening you want. NTRVSTA is also newer and smaller than the large incumbents. Running Greenhouse or Lever instead? See our Greenhouse and Lever guides.
Questions
Does Ashby fraud detection reject candidates automatically?
No. Ashby's fraud detection surfaces signals, and a person marks each candidate as Fraud, Not Fraud or Unsure. Choosing Mark Fraud opens the archive flow with the reason Rejected by Organization: Fraudulent Candidate, so a human makes the call.
Which Ashby plans include Application Limits?
Ashby's docs list Application Limits on Legacy Plus, Plus and Enterprise, not Foundations. Candidate Fraud Detection is listed as available across all plans, and Global Application Questions and Global Auto-Reject Rules are Enterprise only.
Why do Indeed applicants get past our Application Limits?
Ashby matches repeat applicants by email, and its docs say Application Limits do not apply to applications that come through Indeed. Use duplicate detection and your review stages for those.
Does unlisting a job posting stop spam?
Not fully. Ashby notes that unlisted postings can still be scraped and applied to. Only unpublishing a posting stops new applications.
Can an AI screening interview tell us if a candidate is real?
No. Live follow-up questions make relayed or memorized answers harder to keep up and easier for a reviewer to notice, but that is not identity verification. Keep Ashby's fraud review in place and add an ID check from your background-check provider.
Want to see Carolina score a real Ashby pipeline against your scorecard? Book a demo and we'll walk through it with your team.
Sources
- Ashby: Candidate Fraud Detection overview and admin settings
- Ashby: running fraud checks and reviewing fraud signals
- Ashby: Job Board Application Spam Protection
- Ashby: Application Limits
- Ashby: handling multiple applications from one candidate
- Ashby: Auto-Reject Applications
- Ashby: Global Application Questions and Global Auto-Reject Rules
- Ashby: AI-Assisted Application Review
- Ashby: verify answers using AI-assisted analysis
- Ashby: AI features in Ashby
- Ashby: merging duplicate candidate profiles
- Ashby API: applicationSubmit webhook
- Ashby: tools to help identify fraudulent applications
- Ashby: job postings
- LinkedIn: hiring product update (2025)
- LinkedIn: job seeker and HR research (January 2025)
- Greenhouse: AI has doubled recruiters' workloads (November 2024)
- Greenhouse: hiring manager and job seeker survey (November 2025)
- Gartner: candidate survey and fake profile prediction (July 2025)
- FBI IC3: North Korean IT workers (January 2025)




