Auto-anonymisation: what gets removed

#What gets removed or replaced?

#What replaces the candidate's name?

Not initials, and not a blank. The name is replaced with a professional headline summarising seniority, years of experience and specialisation:

This is deliberate. A blank where a name should be tells a hiring manager nothing; a headline tells them immediately whether the candidate is worth reading, without revealing who they are.

#What is deliberately kept?

This is the part that distinguishes a useful anonymised CV from a useless one.

• Quantified achievements. "Managed a team of 10 and a budget of $500,000" survives verbatim. Strip the numbers and you have removed the evidence the client is hiring on.
• Job titles and seniority. The client still needs to know whether this is a manager or a director.
• Date ranges and durations. Career progression is visible.
• Skills, tools and technologies. Industry-standard tools are kept verbatim — Jira, Salesforce, AWS. Only proprietary internal tool names are genericised, since "AcmeCorp Internal Dashboard" identifies the employer.
• Industry context. An employer becomes a short descriptor of industry and type — for example Global fintech company; public. Placeholders like "Company A" are explicitly forbidden.
• Acronyms. KYC, AML, ERP and the like are preserved rather than expanded, because expanding them changes the meaning a specialist reads.

#How are employer names replaced?

With a short descriptor giving industry and type, capped at 12 words. Stripe becomes Global fintech company; public. Size is only added where it is definitively known rather than inferred — if uncertain, it is omitted entirely.

This keeps the CV assessable. "Five years at a tier-one investment bank" tells a hiring manager what they need; "five years at [REDACTED]" tells them nothing.

#Why must I still review the result?

Because identifying information hides in places pattern-matching does not reach:

• Project or product names unique to one employer
• Internal job titles that only one company uses
• Company-specific acronyms in bullet points
• Distinctive career paths — a single named role at a small organisation can be identifying on its own
• Named awards tied to one company
• Acronyms, which are deliberately preserved and can be employer-specific

Automation catches the structured fields reliably. The long tail is a human judgement call, and the consequence of missing one is a candidate identified in what was supposed to be a blind process.

Before anonymising, it is worth confirming nothing was dropped in extraction — see checking what was extracted.

#Can I anonymise a CV I processed earlier?

Yes. Open it from CV history and anonymise it there. No re-upload and no credit — useful when a client switches to a blind process mid-shortlist.

#What if I only want to remove some details?

Edit the fields directly in the editor. Some clients want the employer visible but the name hidden; that is an edit rather than a full anonymisation.

For anything identifying that anonymisation left behind — a project codename, an internal job title — change the wording in the editor before downloading. Replacing beats deleting: "Led delivery of Project Nightingale" → "Led delivery of a group-wide CRM migration" keeps the achievement legible.