Tailoring Is Choosing Your Three Proofs, Not Swapping Keywords

What Changes When You Tailor a Resume

Tailoring used to mean two different things bundled under one word: rewording your resume to match a posting's vocabulary, and choosing which of your real accomplishments to feature for a specific role. The first is mechanical. The second is judgment, and it is the half that actually changes how a reader responds.

At least one major applicant-tracking system's recruiter search is a literal string match: Greenhouse's Talent Filtering states that to appear in a recruiter's results, "the keyword from your search must exactly match the keyword in the application" ¹. That is a real, narrow reason to name a tool or a skill the way a posting names it, when you genuinely have that skill. It is a recruiter looking for people rather than a filter hunting for reasons to reject, and it is not a reason to rewrite every sentence on the page to mirror the posting's phrasing.

Keep the two apart in your own head as you decide how to spend your time. A missing exact term is a small, fixable gap worth ten seconds of attention. A missing accomplishment is not fixable by wording at all, and no amount of rephrasing turns an experience you don't have into one a reader will believe you do.

Why the One-Click Rewrite Stopped Working

It stopped working because everyone got it at once. The builder category automated the mechanical half of tailoring: paste a posting, and a tool rewrites your document to match its vocabulary in seconds. Robert Half, a staffing firm, surveyed more than 2,000 US hiring managers in November 2025 and found 67% saying that reviewing AI-generated applications has slowed their hiring ². That is what employers believe about their own inbox rather than a measurement of any document.

A study of an AI cover-letter tool introduced on a large online labor platform found that access to it raised callback rates at first, and then that the correlation between how well a letter matched a posting and whether it got a callback fell by 51% ³. The study is an unreviewed preprint reading one platform's data rather than an experiment, and the 51% is a fall in that correlation, not a fall in anyone's callback rate. What it describes is a signal losing value as it becomes free to produce: once most documents say roughly the same thing in roughly the posting's own words, matching carries less information, and readers weight whatever still does, including prior work history. What's left to ask for once the cover letter stopped meaning anything is a live question on the employer's side of this same shift, and it is worth reading how they are answering it.

Matching the posting's language is also weaker evidence than it looks in the other direction. In the Harvard Business School and Accenture 'Hidden Workers' survey of 2,275 senior leaders in Germany, the UK and the US, fielded in early 2020, 88% of employers whose organisation runs filtering software said they believed qualified high-skills candidates were vetted out for not matching a job description's exact criteria, rising to 94% for middle-skills roles ⁴. That is an opinion employers hold about their own process, not a rejection rate. It still means an unmatched word is weak evidence that you lack the skill, because it is frequently a phrasing gap, which is exactly the part a rewrite tool cannot see past.

Choose Your Three Proofs Per Application

Keep one honest master resume that lists everything true about you, in full. For a role worth real effort, change three things: the one-line summary at the top, the first bullet under your most relevant experience, and which two or three projects or roles you lead with rather than bury.

That is a selection decision, not a rewriting one, and a model can help you think it through, but it cannot make the choice for you, because it does not know which of your accomplishments you are proudest to defend in an interview. Ask it to summarize the posting's actual priorities in plain language if that helps you see what to lead with, then make the choice yourself.

For roles that don't justify that ten minutes, send the master document as it stands rather than skip the application or force a rewrite you don't have time to do well. No measurement in the sources here shows that tailoring a resume changes callback rates either way, so treat the ten minutes as a bet on one reader's attention rather than on a known return, and note that a rushed rewrite spends it on the half that is now free for everyone. Copying a posting's own words has one narrow, real use, and matching for its own sake is not it.

What Counts as a Proof When Your Background Doesn't Match the Title

A proof is a piece of work that transfers on its own logic rather than one that only sounds related to the new field. If you're changing fields or moving between industries, that distinction carries more weight than it does for someone with a direct-match history, because your three proofs have to do the work a matching job title would otherwise do for free.

A transferable project, ran a budget, managed a launch end to end, resolved a recurring customer problem nobody else would own, taught yourself a tool well enough to ship something real with it, says more about your range than a job title borrowed from an adjacent field. Name what you actually did in the language of the new field once, in the summary line at the top of the page, and let the rest of the resume carry the specifics in plain, ordinary terms rather than forcing every past role into vocabulary it never used and you never worked in.

The ten-minute limit matters most right here, because the temptation to over-explain a career change eats far more time than choosing three lines does, and a long justification reads as uncertainty rather than confidence. State the change once, briefly, in your own plain words, and spend the time you saved on the proof itself rather than on defending why it's on the page at all.

The volume behind this is real where it has been counted: UK graduate employers told the Institute of Student Employers they received an average of 140 applications per vacancy in each of the last two years measured, the highest in the three decades since it began tracking ⁵. That figure covers UK graduate schemes at large employers rather than an ordinary job advert anywhere. Spending equal, shallow effort across every application, tailored or not, is the version of this work that ends up helping nobody, including you.

See a sample report >