Resume tailoring glossary

A proof-gated rewrite is an AI-generated revision to a resume bullet, section, or whole document that has been verified against the facts in the source resume. If the rewrite introduces a metric, tool, certification, company, or job title that the source did not contain, the rewrite is rejected before it ever reaches the user.

Proof-gating treats fabrication as a safety problem, not a stylistic one. An unconstrained LLM will happily invent "increased revenue by 47%" when asked to strengthen a vague bullet: proof-gating refuses to ship that change and instead surfaces the gap so the user can supply real evidence.

ApplySmooth implements proof-gating in three stages: extract facts from the original resume, constrain the rewrite prompt to those facts, and run a separate verifier on the model output. Every rewrite path in the product (bullet, section, full-resume, and Intelligence chat confirmations) passes through the same gate.

The Fix Queue is the operational view of a resume tailoring session. After ApplySmooth analyzes a resume against a job description, it produces a ranked list of suggested changes: weak phrases to rewrite, missing evidence to surface, bullets to reorder, and sections that need more density.

Each item in the Fix Queue is grouped by resume section (Experience, Skills, Education) and ranked by severity. High-severity items are addressed first because they move the match score the most. Medium items are recommended next. Low items are polish.

The Fix Queue exists to remove the question "where do I start?" from a tailoring session. The user works from the top, accepts or skips each suggestion, and watches the match score and Authenticity Check update as they go.

The Authenticity Check is a detective-style score that runs across the whole resume after each Analyze pass. It assigns penalties for patterns that signal embellishment or vagueness, bullets without numbers, bullets with extraordinary claims that lack scope, generic responsibilities phrased as accomplishments, and aggregates them into a single Authenticity score with a per-category breakdown.

The largest penalty weight is reserved for bullets that contain no numbers. Vague bullets without numeric scope are the most common failure mode in resumes and the most common reason a recruiter asks a follow-up the candidate cannot answer.

The Authenticity Check recomputes only on an explicit Analyze action. It does not auto-fire on edit, focus, or scroll. The breakdown is shown to the user so they understand why the score moved.

An Applicant Tracking System is the software layer between a job applicant and the recruiter. It receives the resume file, extracts the text into a structured database, and ranks the candidate against the job description. Most mid-sized and larger employers use one.

An ATS does not care about fonts, colors, or design polish as long as the text underneath is real and selectable. What it cares about is layout. Tables, two-column layouts, text inside headers and footers, and images of text are the four most common reasons parsing scrambles or drops information, including, frequently, the candidate's contact details.

ApplySmooth's free ATS Resume Checker at /ats-score scans an uploaded PDF or DOCX for the specific patterns that ATS parsers struggle with and surfaces them as fixable issues before the resume is submitted.

A match score is the headline number on a Resume Speed tailoring session. It expresses how well the current resume aligns with the pasted job description, on a 0–100 scale, and it updates as the user accepts rewrites and applies fixes from the Fix Queue.

The score is computed from several signals: keyword overlap between the resume and the JD, evidence quality (do bullets contain numbers, scope, and outcomes?), the presence of role-specific tooling the JD calls for, and the Authenticity score that penalizes inflated claims.

A high match score is necessary but not sufficient. ApplySmooth deliberately couples it with the Authenticity Check so a candidate cannot stuff keywords to inflate a number, both scores have to move together for the resume to be a strong submission.

A resume run is the unit of pricing and the unit of work on ApplySmooth. The user uploads a resume, pastes a job description, and the pair becomes a run. Inside the run, the user can re-analyze, accept rewrites, re-order bullets, ask the Intelligence chat for help on specific fixes, and export the result.

Daily run limits separate tiers: Free includes 1 run per day, Starter 12 per day, Pro 30 per day. A purchased run credit expires one year after purchase.

Pricing per run rather than per seat or per token reflects how candidates actually use a resume tailor, one application at a time, in bursts. The user pays only when they would actually send the resume.