AI Resume Builders vs. ATS: What the Research Does and Doesn't Show

January 28, 2026 · FilteredIn Research Team

AI resume builders are often marketed as a way to beat applicant tracking systems. The published research is narrower than the marketing. This page covers the two studies most often cited and what each one can and cannot tell you.

A field experiment on writing assistance

In a field experiment with nearly half a million jobseekers on one online labor platform, those given algorithmic writing assistance were hired about 8% more often (van Inwegen, Munyikwa and Horton). The setting is one platform, the intervention is writing help, and the outcome is hiring on that platform. It did not test ATS optimization or any specific resume tool, and it does not establish effects on callbacks at employers that use an ATS.

A simulation of same-model screening

A 2025-26 simulation in which the same LLM both wrote and screened resumes across 24 occupations reported that the model favored resumes it had written (Xu, Li and Jiang, arXiv). This is a simulated result. It does not show how deployed ATS rank real applicants, and it does not show that applicants can know or match an employer's model.

What vendors document about formatting and parsing

Vendors document the variation. Greenhouse says that when a resume cannot be parsed, it can stay attached to the candidate record and the recruiter enters the details manually (Greenhouse Support). Workday notes that parsing "results can vary based on resume format and order of words" (Workday Documentation). Greenhouse also lets an employer configure automatic rejection based on an applicant's answer to a specific application question (Greenhouse Support). These are examples from two vendors, not a description of every ATS or every employer.

Using an AI builder sensibly

  • Treat generated text as a suggested edit, then verify it and keep it accurate and your own (Harvard Mignone Center).
  • Keep the layout simple and the headings conventional, since parsing results can vary with format.
  • Use the posting's wording where it truthfully describes your experience, not as keyword stuffing.
  • Check each employer's instructions. Some restrict AI-generated content (NIST Careers).

FilteredIn's free resume check lets you paste or upload a resume (PDF or TXT) and, optionally, a job description. It returns feedback on content and keywords. A score is product feedback to help you revise; it is not a prediction of any employer's ATS or hiring decision. Try the free resume check.

Sources