For employers

A CV is the smallest possible description of a person.

It is also the only thing your screening reads. So a language it cannot parse, a title it does not know, and a career that changed direction all come out the same way: no. The competence was there. The description was not.

And first-year turnover is eating the hires you do make. You find out when the resignation arrives.

Request a pilot conversation See the three tiers

The one thing to know first

The platform makes no hiring decisions.

It ranks, it explains, and it stops. No automatic rejection, no auto-advance, and no ranking whose basis your recruiter cannot see.

Every shortlist arrives with an explanation attached, and every human decision is an explicit action logged against the deciding user’s identity and exportable for labour inspection.

This is not a caveat we added for the compliance page. Employment-related matching is treated under the EU AI Act’s high-risk obligations, and human oversight is one of them. It is a constraint in the product, not a setting you can turn off.

Foreign nationals working

3.1M

In the Spanish economy.

SOURCE: INE AND SOCIAL SECURITY · 2024

Of new jobs

79%

Share of new Spanish jobs filled by migrant workers in 2024.

SOURCE: OECD INTERNATIONAL MIGRATION OUTLOOK · 2024

New permanent immigrants

368,000

Admitted to Spain in 2024, fifth in the OECD.

SOURCE: OECD INTERNATIONAL MIGRATION OUTLOOK · 2024

Where the loss happens

Four ways a good candidate disappears.

None of them is a judgement anyone made. All four happen before a human reads a word.

What we describe instead

Six fields, five of which only the candidate can fill.

A person is not a title and a keyword list. They are what they can do, the languages they have, the shape their career took, how they work with people, and what they want the work to be for.

None of these six is a filter you can select on.

A role declares what it needs. That is matched against what a candidate has stated, and the recruiter reads why. Nothing that maps to a protected characteristic is a ranking input, and no combination of fields reconstructs one. The constraint is in the query layer rather than in the interface, which is the difference between a rule and a preference.

This is also why the platform is not a diversity screening tool and will not be sold as one. Cohort composition is reportable in aggregate, above the minimum cell size, for the employer’s own reporting obligations. It is never available as a selection criterion on an individual, and there is no configuration that makes it one.

One profile, completed once, that is not a job application. It asks what they can do, what they have done, and what they want the work to be for. It is additive to a CV rather than a retyping of one, it is not redone per employer, and nothing in it is required. An incomplete profile is shown as incomplete rather than scored down.

The gate

Your recruiter meets a competence profile. The person comes later.

Screening bias does its work before anyone has read a competence: on a name, a photograph, a birth year, a postcode. So the shortlist does not carry any of them.

Identity is released when a named recruiter records an intent to hire against one role. That action is logged, and it is the only thing that opens it.

Sealed on the shortlist

Name, and any photograph
Date of birth, and any date that implies it
Nationality, country of birth, and residence status
Address, postcode, and neighbourhood
Gender
The names of previous employers
The name of the awarding institution, though not the qualification or its level

Shown on the shortlist

Competences, each with its evidence and its confidence value
Which competences are adjacent, and which are missing
Languages and levels
Years of practice, by competence rather than by employer
Everything the candidate wrote about themselves

Intent to hire is an explicit action taken by a named recruiter against one role. It releases the identity to that recruiter, for that role, and it is written to the audit log with their identity on it and exported for labour inspection. There is no bulk reveal, no reveal by export, and no setting that turns the gate off.

What is not in here

There is no feed, because a feed decides who you see.

A network that ranks people by how much they use it is measuring availability for the network. These three are absent from the data model, not switched off in a setting.

A candidate who has never opened the platform ranks exactly as well as one who checks it daily. Recency of activity is not a field, so it cannot be a tiebreak.

There is no promoted placement and no candidate tier. Employers pay for seats and modules, and no amount of either moves a person up a list.

Connections, endorsements, followers, and post engagement are not collected. A network that ranks people by how much they use it is measuring availability for the network, not suitability for the work.

Module 01

Workforce Matching

Matching on a description of the whole person rather than on a title: competence against the European ESCO framework, every language they have, the shape of the career, and their own account of the work they want. The shortlist names nobody until a recruiter records an intent to hire.

What your recruiter sees

Not a score. A ranked list, and for each candidate: which required competences matched and on what evidence, which are adjacent, which are missing, and the confidence value. Then they decide.

Shortlist · explanation panel · anonymous

Sample data

Candidate: CAND-4471 · identity sealed until an intent to hire is recorded
Matched: Installs electrical systems · Reads technical drawings · Tests installations · Applies safety regulations
Adjacent: Planned maintenance scheduling, evidenced in residential rather than industrial context
Missing: Industrial control systems
Confidence: 0.82 · above threshold
Decision: Awaiting a named recruiter. The platform does not advance this candidate.

How ESCO alignment actually works, and where it breaks

Bilingual by construction

The models read Spanish and Catalan natively rather than through English, version-pinned, with the model version recorded on every output. Not a translation layer over an English model. Language is a first-class field on every record, recorded at ingestion rather than inferred at query time.

Show the data table

Approach	Qualified candidates recovered, indexed
Keyword scan	34
Translate then match	61
Native ESCO match	84

Illustrative comparison of matching approaches on a bilingual corpus. These are internal evaluation figures on a test set, not results from a client deployment, and they are not comparable to anything you can independently check. The first employer engagement will produce measured figures.SOURCE: HUMANIWORK INTERNAL EVALUATION SET · 2026 · MODELLED

Module 05

Employer Intelligence

Stability and attrition-risk profiles at cohort level, plus sector concentration analytics. Tells an employer where retention will break before it breaks.

What it sees

Cohort-level retention patterns across roles, sites, and sectors. Never an individual worker.

What you do with it

An HR director sees which role and site combinations are heading for a retention break, and moves before the resignations arrive.

The floor

Cohorts below the minimum cell size are suppressed at query time across every filter combination, including exports.

There is no individual-level attrition score, and no combination of filters produces one. That is enforced in the query layer, not in the interface.

Built for the DPO

Compliance is the doorway, not the disclaimer.

Your labour inspector and your works council ask the same questions. These are the answers.

Anonymisation at ingestion
No individual-level data reaches the AI layer.
No individual-level data enters the AI processing layer. Analysis runs on aggregated and pseudonymised data, and the rule is enforced at the pipeline boundary rather than by convention.
EU data residency
EU infrastructure, Catalan residency for Catalan deployments.
Infrastructure in EU jurisdiction, with Catalan data residency for Catalan deployments. The sub-processor list is published and versioned.
Per-module EU AI Act classification
Classified per module, before deployment, because they differ.
The classification is documented per module before deployment, because the modules are not equivalent. Employment-related matching is treated under the high-risk obligations. The civic analysis modules are advisory decision-support for institutions.
A human decides
The platform ranks and explains. A person decides.
The platform ranks, explains, and projects. A qualified human makes every hiring and every policy decision, and the decision is logged against their identity.
DPA before ingestion
Signed, and your DPO briefed, before a record moves.
A signed Data Processing Agreement and a briefed Data Protection Officer before a single record moves. The system refuses to create a data source without a recorded DPA reference.
Exportable audit trail
Every output keeps what produced it, and exports.
Every match, alert, and projection persists its inputs, model version, driving features, and confidence value. Exportable for labour inspection and public audit.

EU AI Act classification, per module

Module	Classification	What that means in practice
01 Workforce Matching	High-risk obligations	Employment-related matching is treated under the high-risk obligations: risk management, data governance, logging, human oversight, transparency, and accuracy documentation. The classification and its reasoning are versioned in the repository and ship with every release.
02 Social Observatory	Advisory decision-support	Civic analysis is advisory decision-support for an institution. It produces no individual-level output and no automated decision. The classification and its reasoning are versioned in the repository.
03 Integration Intelligence	Advisory decision-support	Advisory decision-support at cohort level. No individual-level output, no automated decision. Classification versioned in the repository.
04 Policy Simulator	Advisory decision-support	Simulation output is advisory decision-support and is labelled modelled everywhere it appears. Classification versioned in the repository.
05 Employer Intelligence	Advisory decision-support	Cohort-level analytics with an aggregation floor applied at query time. It produces no individual-level risk score and cannot be filtered down to one. Classification versioned in the repository.

Source: Humaniwork compliance repository, versioned per release, 2026

The full compliance surface, including the sub-processor list and the data flow

What you keep

You keep everything.

Whatever we build from your data, you keep. The raw files, the derived datasets, the methodology, and the platform access. If you stopped working with us tomorrow, you would still hold everything we produced together.

A consultancy keeps the model

A software vendor keeps the data

You keep both

What you receive under the Handover Standard, in what format, and when.
What you get	Format	When
Every raw and derived dataset we build from your data	CSV or Excel, with a data dictionary	Rolling, from the first pipeline run
The indicator and index definitions, in full	Written methodology document	Diagnostic phase, before anything is computed
Data pipeline documentation and runbook	Technical PDF and configuration	On pipeline delivery
Every intelligence brief and report	PDF, bilingual as standard	On the delivery cadence
Model parameters and configuration	Documented and transferred to your records	Handover phase
Platform access for your named users	Live, role-based	From setup onward
Knowledge transfer workshop	Half a day, with your team	Handover phase

And, for this engagement specifically

The audit trail. Exportable, complete, and yours. For a buyer facing AI Act obligations it is arguably the product.

Every dataset, document, and definition above is already yours and stays yours. What is still being settled is how long platform access continues after an engagement ends and is not renewed. Ask in the diagnostic conversation and you will get the current answer in writing rather than a shrug.

Pricing

Three tiers, with the figures on the page.

Annual, all-inclusive, onboarding included. No discounts, no bundles, no limited-time pricing.

Essential

€12,000

Per year. Single site or single hiring function.

Workforce Matching in Spanish and Catalan, ESCO-aligned
Explanation attached to every shortlist
Exportable audit trail
Up to 3 recruiter seats

Professional

€20,000

Per year. Multi-site, multi-function.

Everything in Essential
Employer Intelligence: cohort-level stability and attrition-risk profiles
Sector concentration analytics
Bias-monitoring reporting on every matching run
Up to 10 recruiter seats

Enterprise

€30,000

Per year. Group-level, with identity federation.

Everything in Professional
OIDC or SAML federation with your identity provider
Custom ESCO mapping for internal role taxonomies
Named analyst and quarterly methodology review
Unlimited recruiter seats

All figures are annual and all-inclusive. Onboarding is included. There are no discounts, no bundles, and no limited-time pricing.

Reviewed 2026-01 · EUR

Start with a diagnostic.

Two weeks, before any commitment: what you are trying to fix, what your hiring data can currently support, and an honest account of what the matching engine will and will not do on your vacancies.

Request a pilot conversation The compliance surface, for your labour inspector How an engagement runs