The fastest test is to read the bullets back aloud. Raw ChatGPT bullets sound like a competent press release about a fictional person. They use words like "spearheaded", "leveraged" and "drove strategic initiatives". They feature metrics like "increased revenue by 47 percent" that look clean but were not in your inputs. They flow from one to the next with identical structure.
Hiring managers see this pattern across the pile. The 2024 ResumeBuilder survey put detection confidence at 74 percent (the TopResume blind test puts actual accuracy at 33.5 percent, which is still well above chance). The 2026 data on outcomes: 49 percent of US hiring managers auto-dismiss suspected AI CVs, 62 percent reject AI CVs that lack personalisation.
The technical problem is ATS scoring. Raw output averages 68 out of 100 on ATS compatibility checks. Most systems treat 80 as the human-review threshold. Below 80, the parser sees missing keywords, missing standard headings, formatting guesses that do not match what Workday or Greenhouse expect. The fix list is published in detail in our 30-minute optimisation workflow.