Practice lines How it works Research About us Request an engagement ES · Español
Data · Risk · Sustainability

Stop reporting on social risks.
Start architecting resilience.

Movaterra helps organisations quantify, predict, and assure the human side of risk across their workforce, their AI systems, and their supply chains.

Request an engagement → What we deliver
Why now

Social risk stopped being a reporting exercise.
It's now a legal exposure.

AI systems used in employment are classified high-risk under the EU AI Act; the 2026 Digital Omnibus pushed most Annex III high-risk obligations to December 2027, but transparency duties are live now and fines of up to €35M or 7% of global turnover still apply once obligations bite. In New York, Local Law 144 requires employers to run an independent bias audit every year before using an automated tool to screen or assess candidates, and Colorado's automated decision-making technology (ADMT) law follows in January 2027. In Europe, the Corporate Sustainability Due Diligence Directive and the EU Forced Labour Regulation are turning human rights due diligence in the supply chain into something regulators and courts can actually enforce. Wherever you look, the question is the same: can you show your evidence? Movaterra exists to make that evidence rigorous, quantitative, and defensible.

Practice lines

Three ways to engage.

All remote, all defined-scope, all delivered on the Movaterra engine by the researchers who built the methods. We compute, we report, and we deliver training on our methods; we don't facilitate open-ended change programmes.

Assurance

Assurance

Independent audit for consequential AI decisions about people, benchmarked against ISO/IEC 42001 and 42005: hiring, screening, monitoring, scheduling, performance.

  • Risk Classification & Gap Assessment. System inventory, EU AI Act classification, a gap analysis benchmarked against ISO/IEC 42001 (AI management systems) and Chapter III, and a board-ready remediation roadmap.
  • Independent Bias & Worker-Impact Audit. Quantitative fairness testing (selection rates, impact ratios, counterfactual analysis) combined with a worker-impact assessment structured on ISO/IEC 42005 impact-assessment methodology and fundamental-rights impact logic, the dimension no algorithm-only auditor covers.
  • Conformity Retainer. Post-market monitoring design, re-audit triggers, and documentation upkeep aligned to your AI management system.
Analytics

Analytics

Predictive social risk for leaders who need to know where to look before the auditor, regulator, or journalist does.

  • Supply Chain Social Risk Prioritisation. Predictive modelling that ranks suppliers, sites, and corridors by forced-labour and labour-rights risk, directing deep-dive investigation where static audits fail. Built for CSDDD, EU FLR, and UFLPA exposure.
  • Risk Analytics. Custom predictive risk modelling built on your own data, the same engagement model behind the UN's workforce risk tool: predictive modelling and counterfactual attribution applied to the social risk your organisation needs to track.
  • Social ("S") Metrics & Disclosure Support. Investor-grade measurement design for the least-measured pillar of ESG.
Academy

Academy

Executive education for boards and senior leaders navigating workforce AI risk and digital transformation.

  • Board briefings on workforce AI risk and regulation.
  • EU AI Act Article 4 literacy programmes.
  • ISO/IEC 42001 & 42005 readiness training for governance, risk, and compliance teams, ahead of third-party certification.
  • Executive education on digital leadership, digital strategy, and digital transformation.
Independence

We don't audit what we built.

Assurance and advisory are kept strictly separate. Movaterra will not audit any system it has helped design, build, or remediate. If advisory work would create a conflict for a future audit, or the reverse, we disclose it and decline the conflicting engagement. Audit opinions are issued under the name of the responsible auditor and are never contingent on advisory revenue.

How it works

Your data in. Defensible evidence out.

Every engagement follows the same pipeline, whether the output is an audit opinion or a predictive model, running on our own engine or on your systems. That's why the second wave is always faster than the first, and your results stay comparable across years.

01 · scope

Scope

Defined scope, instruments, deliverables, and timeline, agreed before we start. You know exactly what you get.

02 · collect

Collect

We deploy validated surveys or receive your data through secure channels. Anonymity thresholds are enforced in the engine, as code, not as a promise.

03 · quantify

Quantify

Statistically validated scoring produces composite indices with sub-factor breakdowns by unit, site, and cohort. ML risk modelling where longitudinal data allows.

04 · deliver

Deliver

An interactive dashboard plus board- and regulator-ready reports. Re-runnable as a measurement wave: evidence with a date on it.

One engine, every engagement. All delivery runs on the Movaterra engine: the same tested codebase applying peer-reviewed, published methods, growing sharper with every project. Our founder designed the United Nations' first predictive workforce risk model; the same discipline runs through everything we take on.

Our research →
Why Movaterra

Rigour without compromise.

Our methods are grounded in over a decade of peer-reviewed science, funded by leading research councils and deployed at international institutional scale. We publish what we build, and we build only what the evidence supports.

Machine learning, not just metrics
Our toolkit is predictive modelling: gradient boosting, NLP, counterfactual attribution, and advanced feature construction, applied where it changes decisions, not where it decorates reports.
Published, not proprietary
The science behind our work is peer-reviewed and public. Anyone can check the methods: that is the point.
Regulatory-grade by design
Everything we build is structured against ISO/IEC 42001 (AI management systems), ISO/IEC 42005 (AI impact assessment), and social responsibility standards, designed to stand up to regulators, auditors, and investor due diligence.
Privacy as a feature
Anonymity thresholds, lawful data handling, and accredited-researcher discipline are built into the measurement itself, not bolted on afterwards.
Production ML, not a one-off
Full-stack machine learning practice: the same tooling and discipline behind the UN model, applied to every engagement.
Research

The science behind our methods.

Movaterra's methods are built on peer-reviewed research authored by our founder and collaborators: published, public, and open to scrutiny. A selection below; the full record lives on ORCID and Google Scholar.

Futures of Work · 2026 · Commentary

From rates to risk: how machine learning can reveal the workers official statistics cannot see

Applying machine learning to underemployment data to surface the workers invisible to headline labour statistics.

Read article → Safety Science · 2023

The potential of responsible business to promote sustainable work

An analysis of CSR and ESG instruments and their capacity to drive genuine sustainability in the workforce.

Read paper → Social Science & Medicine · 2022

The impact of national legislation on psychosocial risks

Quantitative analysis of how European legislation shapes organisational action plans, working conditions, and work-related stress.

Read paper → Sociology · 2025

Carrying the domestic burden of the Covid-19 pandemic

Gender, class, and the domestic division of labour through turbulent times.

Read paper → British Journal of Sociology · 2024

Class, gender and the work of working-class women

Intersecting inequalities in UK working lives amid economic turbulence.

Read paper → Springer · 2021 · Edited book

Aligning perspectives in gender mainstreaming

Gender, health, safety and wellbeing across organisational and policy contexts.

View book → Safety Science · 2017

Employer's civil liability for work-related accidents

A comparison of non-economic loss in Chile and England.

Read paper → UK Parliament · 2024 · Policy

Evidence on the Employment Rights Bill

Written evidence to the Business and Trade Committee, from the ESRC underemployment research programme.

Read evidence → The Underemployment Project · 2023 · Data

Underemployment levels and trends

An open data portal mapping time-, skills-, and wage-based underemployment in the UK.

Explore the data →

Looking for the full record? Every publication, with citations, is on our founder's public profiles.

Google Scholar → ORCID →
About us

One question,
four disciplines.

Movaterra exists to answer one question: how do you make an organisation's impact on people measurable, rigorously enough to change decisions?

Movaterra works across four disciplines: management science to understand organisations; socio-economic, survey, and organisational data to see them clearly; sustainability and compliance expertise to know what matters; and machine learning to make it all operational. Every engagement is delivered by the researchers who built the methods.

The people behind Movaterra
Luis D. Torres
Luis D. Torres, PhD
Founder · Associate Professor, Nottingham University Business School · Associate Professor, Universidad del Desarrollo, Chile · Data science consultant, United Nations

Luis designed the United Nations' first predictive workforce wellbeing risk model, built originally at UNHCR and now scaling across multiple UN agencies. He holds a PhD in Management from the University of Nottingham, where he is Associate Professor and previously served as Director of Executive Programmes, and postgraduate qualifications in environmental law and international human rights law. He is also Associate Professor at Universidad del Desarrollo in Chile and a Fellow of the Institute of Corporate Responsibility and Sustainability. He sits on the ISO/TC 309 working group on human trafficking, forced labour and modern slavery, contributed to the British Standard BS 25700 on organisational responses to modern slavery, and serves on the editorial board of Safety Science. He is a Full Accredited Researcher with the UK Statistics Authority and builds the predictive models himself rather than commissioning them.

LinkedIn Google Scholar ORCID
Start here

Request an engagement.

Tell us what you need and what data you have, or don't have; deploying the instruments is our job. You'll receive a scoped proposal covering instrument, timeline, and deliverables.

Sense the risk.
Prove the return.

Independent AI audits, predictive social risk analytics, and executive education, delivered.

Request an engagement →