Workable CV tips: writing for a semantic AI screener (2026)

Workable sits on thousands of small and mid-sized companies' hiring, and its AI screening reads meaning, not just keywords.

Workable parses your CV into a candidate profile and, as of August 2026, adds AI screening: semantic matching of your experience against the job requirements, plus AI-generated summaries for the hiring team. Semantic means related terms count, so write clear, quantified evidence of each requirement rather than stuffing exact keywords.

How Workable handles your CV

Workable is a widely used ATS among small and mid-sized companies worldwide. Applying is light: a short form, CV upload, usually no account. Your CV is parsed to populate a structured candidate profile the hiring team works from.

As of August 2026, Workable's screening layer is AI-assisted: it matches candidates against the role's requirements semantically, surfacing skill and experience gaps at a glance, and generates profile summaries so recruiters do not read every CV line by line. Semantic matching means "led the sales team" and "sales management" connect without identical wording.

Decisions stay human: the AI orders and summarises the pipeline, and employer-configured screening questions can filter, but there is no autonomous rejection of CVs.

What to do: the Workable checklist

  • Write one clear, quantified evidence line per requirement in the advert: Workable's semantic matcher connects related phrasing, so substance surfaces you and stuffing does not.
  • Quantify everything you can: AI summaries compress your CV for the recruiter, and concrete numbers are what survive compression.
  • Keep the profile-critical facts (current title, location, latest role) unambiguous in the top third of the CV, where the parse anchors your profile.
  • Answer screening questions carefully; they remain the only automatic filter in the flow.
  • Use a single-column layout with a real text layer so the parse behind the AI is complete.
  • Skip the exact-keyword paranoia, but keep your terms conventional: semantic matching handles variants, not private jargon.

Myth vs reality

The myth

You must copy the advert's keywords exactly or Workable's AI cannot see you.

The reality

Workable's matching is semantic: it connects "managed a team of 12" with "people management" without identical wording. Exact-copying keywords adds little, and padding your CV with terms you cannot evidence backfires twice, once with the AI summary that surfaces the gap, and again with the human who interviews you.

How Vouch prepares your CV for Workable

Vouch's anti-fabrication design fits AI-screened pipelines like Workable's: every claim in your CV comes from what you actually told Maya in your interview, and a deterministic guard strips terms you cannot honestly claim. Semantic screeners reward exactly that, real evidence in clear language, over keyword decoration.

Every CV Vouch exports is a single-column layout with a real text layer, generated from structured data rather than a screenshot, so any parser in this guide can read it. Start with the free cover letter generator, and if you are converting a US resume for UK employers, see UK CV vs American resume.

Frequently asked questions

Does Workable auto-reject CVs?

No. As of August 2026 Workable's AI screens, ranks and summarises candidates against the job's requirements, but does not reject anyone autonomously; that stays with the hiring team. Employer-configured screening questions are the one automatic filter. A weak semantic match means you rank lower in the pipeline, not that you are removed from it.

How does Workable's AI screening actually read my CV?

It parses your CV into a structured profile, then matches your experience semantically against the role's requirements, flagging where you meet them and where gaps appear, and writes a short summary for the recruiter. Related wording counts: you do not need the advert's exact phrases, you need clear evidence of the underlying skill.

What CV format works best for Workable?

A single-column PDF or Word document with standard headings, your current title and location clear at the top, and quantified achievement bullets. The AI layer can only work with what the parser extracted, so a complete text layer matters; beyond that, optimise for a busy human skimming an AI summary: numbers, outcomes, plain language.

Let Maya write it with you

Vouch interviews you first, in your browser or by phone, then writes applications that sound like you. Free to start.

Start free