The HR partner
who actually works strategically.

Employment references in 10 minutes. 200 applications sorted in two hours. Discrimination-free, compliant with employee consultation requirements.

anymize is the AI workplace for HR departments that want to use GPT, Claude and Gemini – without handing employee data to US providers and without skirting GDPR employee data protection, employee representation requirements or anti-discrimination law. The bidirectional anonymization strips names, salaries, dates of birth, health data and religious information before a prompt leaves the building. For internal work with spreadsheets, org charts and compensation data, our own European-hosted model is ready.

The compliance frameworkStart for free

The promise

Who you
can become.

The HR business partner who finally works strategically – because employment references, job descriptions and first-pass screening stop eating the whole day. The recruiting manager who gets done in two hours what used to take two weeks. The department candidates talk about: “They got back to me within 24 hours – with a qualified answer.”

The tools exist. Claude drafts employment references in the precise legal code language. GPT understands what a job posting needs to look like to be anti-discrimination-compliant. Gemini pulls the relevant qualifications out of 200 CV PDFs. But HR has not used these tools, because it was not allowed to – employee data, health data from medical files, salary information and applicant data.

anymize removes that block. You upload personnel files, reference drafts, application PDFs. The anonymization pipeline replaces names, dates of birth, religious affiliation, salaries and addresses with semantic placeholders. The frontier model responds with full context; the re-translation gives you the result with the original data. From the AI provider's point of view, no employee and no applicant has ever been visible.

The hard compliance edges

What HR has to observe —
and how we address it.

HR AI use touches four regulatory layers across the EU. We have addressed each of them on the product side — the legal final responsibility stays with your organization.

GDPR Art. 6 & 9

Employee data requires explicit legal basis

Processing employee and candidate data under GDPR requires a clear legal basis for each purpose — hiring decisions, performance reviews, payroll, disciplinary processes. Special category data (health, disability, union membership) requires even stricter justification under Art. 9. Any AI tool involved in these processes must stay within the documented legal basis.

What already applies: The GDPR principles (purpose limitation, necessity, data minimization) apply in full. Fully automated individual decisions under Art. 22 GDPR (e.g. a fully automated applicant rejection without human review) require specific legal bases.

How anymize addresses it

  • Purpose limitation through separate projects and workspaces per HR use case
  • Data minimization via the anonymization pipeline — frontier models never see personal identifiers
  • Logging of decision support for evidentiary purposes (audit log)

EU AI Act Art. 26

AI in recruitment is high-risk AI

The EU AI Act explicitly classifies AI systems used in employment decisions — recruitment, promotion, performance evaluation — as high-risk AI. Employers deploying such tools must provide transparency notices to workers, ensure meaningful human oversight, and maintain audit documentation. This applies to any AI tool that influences hiring or evaluation outcomes.

What anymize recommends

  • Sign a works agreement or employee data agreement before rollout; we provide a template with typical clauses (permitted use cases, mandatory anonymization, employee-representative access to the audit log, exit clause)
  • Transparency features: audit log also readable by employee representatives; clear definitions of permitted use cases in the admin area

Equal Treatment & Anti-Discrimination

AI must not introduce bias

Anti-discrimination law across all EU member states — implementing the EU Equal Treatment Directive — prohibits algorithmic bias in hiring and promotion. GDPR Art. 22 further restricts purely automated decisions affecting individuals. Using AI for CV screening or interview scoring requires human review checkpoints and documented bias-mitigation measures.

The central risk: AI models can be trained on historical data that carries systematic biases. Applicant screening tools have produced high-profile anti-discrimination precedents internationally.

anymize safeguards

  • Anonymization removes protected attributes (names that reveal origin; date of birth, which would disclose age; addresses; religious indicators)
  • Bias awareness in prompt design: suggested HR prompts are written so that bias-prone instructions (e.g. "find young, dynamic candidates") stand out and can be blocked
  • Human final decision: anymize is decision support, not a decision automaton. Fully automated individual decisions (GDPR Art. 22) through anymize are not supported by design

GDPR Art. 22 & Automated Decisions

Human oversight in AI-driven HR

GDPR Art. 22 gives employees the right not to be subject to solely automated decisions that produce significant effects. HR teams using AI for screening, scoring, or ranking must ensure a human is meaningfully involved in the final decision. Anonymization of candidate data before AI scoring also reduces the risk of inadvertent discrimination.

anymize itself is not a high-risk AI system — you as the deployer decide how you use the models provided through anymize. High-risk HR use cases (automated pre-screening, promotion scoring) require additional governance.

What anymize provides

  • A Data Processing Agreement (DPA) under GDPR Art. 28 — required for any AI processor handling employee or candidate data, effective automatically on account creation
  • Anonymized AI processing — candidate and employee PII is stripped before reaching AI models, reducing GDPR Art. 22 automated decision risk and bias exposure significantly
  • Audit logs for every data processing event — supporting GDPR accountability requirements and EU AI Act documentation obligations for high-risk AI in employment

Five everyday scenarios

Concrete workflows,
concrete toolchains.

01

Drafting employment references

Toolchain

  • Upload (anonymized): personnel file, performance feedback, prior references
  • Claude Deep
  • Knowledge base “reference standards” active
  • “Draft a benevolent interim reference with an overall grade of good; factor in the last three projects”

Result

Reference in your house tone as an artifact, editable, export to Word with original names

02

First-pass application screening

Toolchain

  • Upload (anonymized) of 200 application PDFs in batch
  • Gemini (multimodal) for CVs with photos (photo is anonymized)
  • “Rank by fit to the job description, ignore name, age, origin”

Result

Ranked long list with a reason per candidate, anti-discrimination-clean

03

Drafting a job posting

Toolchain

  • Project “Open role Senior DevOps”
  • Claude
  • Knowledge base “in-house compensation data + equal-treatment guide”
  • “Write the posting free of discrimination, including all legally required notices under equal-treatment regulations and inclusive-hiring obligations”

Result

Equal-treatment and inclusive-hiring-compliant job posting as an artifact

04

Employee communication (termination, warning)

Toolchain

  • Project “Personnel file Ms. X” (with anonymized file background)
  • Claude
  • “Draft the formal warning with reference to the contractual duty of performance”

Result

Draft warning as an artifact, legal compliance reviewable

05

Internal HR analytics

Toolchain

  • anymize Fountain (internal, no anonymization needed)
  • Upload of aggregated HR metrics (attrition, absences, salary anomalies, without personal reference)
  • “What patterns do you see?”

Result

Analysis artifact, no data leaves the building

Every workflow works today. No “this is coming in Q4” — the toolset is live.

Model recommendations

Which model
for which HR task.

Employment references, terminations, formal warnings

Claude (Deep)

Precise legal-code language, employment-law-sensitive phrasing

Application review with CV PDFs and references

Gemini

Multimodal, strong on CV layouts and PDF structures

Job postings, communication with candidates

GPT (Smart)

Modern, accessible language, strong for employer branding

Research on collective bargaining agreements, case law, industry benchmarks

Perplexity

Web-native research with defensible sources

Very large application stacks (over 100 PDFs)

Kimi (Deep)

Very large context window, all applications in one pass

Internal HR analytics on spreadsheets

anymize Fountain

European-hosted inference, no anonymization needed for aggregated data

Strategic workforce planning, multi-step reasoning

anymize Waterfall

European-hosted inference, frontier-grade reasoning depth

Details on all international modelsDetails on Waterfall & Fountain

Feature set for HR

The tools that land
in the HR department.

Five features with the highest leverage in HR's day-to-day:

01

Knowledge bases

Your own knowledge bases for works agreements, in-house reference standards, compensation data, equal-treatment guidelines, collective bargaining agreements. Every HR team member works on the same, current rulebook.

02

Projects

One project per recruiting process or sensitive personnel file. Context grows with every chat. Colleagues can seamlessly take over during sickness or vacation — the context is documented and searchable.

03

Live transcription

Record interview conversations, employee talks (with consent), return-to-work talks on your phone; transcribed on our own European infrastructure. Massively reduces note-taking — important for low-performance conversations or return-to-work meetings.

04

Artifacts

References, job postings, termination letters, coaching briefs emerge as artifacts — WYSIWYG-editable, exportable to Word or PDF. Export with original names is one click away.

05

Compliance controls

Admin controls: mandatory anonymization (not overridable by individual users), employee-representative access to the audit log, defined permitted use cases, review countdown for sensitive decisions.

Integration into existing HR software

Where anymize
plugs in.

The anymize API is OpenAI-compatible (drop-in replacement) — any HRIS or recruiting automation that can speak to OpenAI today also works with anymize:

01

Automation platforms

Zapier, Make.com, n8n with native anymize nodes: pull applications from the ATS, pre-sort anonymized, write the summary back.

02

HRIS integration via connector

Personio, SAP SuccessFactors, Workday, rexx systems, HRworks, Sage HR, Umantis via API + low-code tools.

03

ATS integration

Greenhouse, Lever, Teamtailor, Workable, Softgarden, Softworx.

04

Email and document pipelines

Outlook automations (e.g. “incoming application with attachment → pre-sort anonymized → structured summary back into the ATS”).

Concrete integration guides: n8n · Make.com · Zapier · MCP Server for Claude Desktop.

On-premise relevance

When employee representatives
don't want cloud.

For organizations whose employee representatives reject a pure cloud solution, we offer on-premise deployments: the anonymization pipeline and our small model anymize Spring inside your own infrastructure. Hybrid variant: local anonymization, anonymized requests to international frontier models — no employee data leaves the company data center.

For most mid-sized HR departments, our cloud solution is enough — with our template employee data agreement and data processing agreement (DPA) under GDPR Art. 28, the path to employee-representative approval is short.

Which plan fits

From one-person HR
to the enterprise group.

Company size

Small HR (< 10 people in HR)

Recommended plan

Professional or Team Business

Why

Full feature set, simple administration

Company size

Mid-sized HR team (10–50 people)

Recommended plan

Team Business

Why

Shared knowledge bases, admin controls, employee-representative access to the audit log

Company size

Enterprise HR (> 50 people)

Recommended plan

Enterprise

Why

SSO, individual employee-data-agreement support, dedicated account manager, on-premise option

Company size

Recruiting agency

Recommended plan

Team Business

Why

Projects per client mandate, knowledge bases per industry, clean mandate separation

All plan details

How you start in three steps

From kickoff
to productive use.

1

Step 01

Create the account, prepare the employee data agreement

Registration concludes the data processing agreement under GDPR Art. 28. We provide a template employee data agreement with typical clauses (permitted use cases, mandatory anonymization, employee-representative access to the audit log, exit clause). Alignment with employee representatives typically takes 2–4 weeks, depending on company structure.

2

Step 02

Pilot with one HR use case

Start with a clearly scoped use case (e.g. drafting interim references, anti-discrimination-checking job postings). First knowledge base: reference standards and in-house compensation data. Define first workflows.

3

Step 03

Rollout, EU AI Act documentation, transparency toward employees

After a successful pilot: roll out further use cases. Admin policies (mandatory anonymization not overridable, four-eyes review for HR decisions). Transparency notices to the workforce pursuant to EU AI Act Art. 26 — we provide a template. The audit log supplies the documentation base for the EU AI Act and GDPR employee data processing.

For whom inside HR

Six roles,
six focal points.

01 · Role

HR business partner

Strategic work instead of operational reference grind; employee coaching with solid preparation

02 · Role

Recruiter

200 applications structurally sorted in two hours; anti-discrimination-clean job postings

03 · Role

HR administrator

Employment contracts, references, certificates in minutes instead of hours

04 · Role

Head of HR / HR director

Audit log as governance basis, workforce planning with anymize Waterfall

05 · Role

Compliance / data protection

GDPR employee data and EU AI Act documentation in one tool, anonymization as a technical measure

06 · Role

Employee representatives / works council

Access to the audit log, transparency on permitted use cases, employee data agreement as control basis

What HR teams often ask.

Frequently asked questions

Usually yes. Employee representation requirements across the EU typically mandate consultation when technical systems capable of capturing employee conduct or performance are introduced. AI-supported HR tools typically fall within scope. We provide a template works agreement or employee data agreement with typical clauses. That speeds up alignment with your works council or employee representatives.

Legal note

The statements on this page on GDPR (Art. 6, 9, 22, 28), the EU AI Act (Reg. 2024/1689), the EU Equal Treatment Directive and related anti-discrimination law, and employee consultation and representation requirements are a classification based on the currently applicable texts (as of April 2026; sources: eur-lex.europa.eu and national implementing laws). They do not replace individual employment-law advice or concrete coordination with your employee representatives and data protection officer. For a binding assessment of your specific situation, please consult a specialized employment lawyer, your data protection officer and your works council or employee representatives.

Start now.
14 days free trial.

All models. All features. No credit card.

Start for freeHow it works

We stand behind anymize. And we know – when an AI tool touches client, patient or employee data, a demo video isn't enough. That's why we give you 14 days of full access – all models, all features, no credit card. Enough time to be certain, before you trust us.

Your AI workplace awaits.