Horizon Europe
AdriaMont Consulting is a research, evaluation and assessment company in Podgorica, Montenegro, a Horizon Europe Associated Country. We join consortia that already have a coordinator, as a partner and as leader of the evaluation and measurement work package, or of a task inside it.
Currently a partner and work-package lead in a Horizon Europe Cluster 2 proposal now under evaluation.
Our subject is what happens to people when AI enters their work: how they experience it, which skills change, whether the system treats groups fairly, and whether the project delivered what it promised.
Study and evaluation design, theory of change, indicator and KPI frameworks, monitoring and evaluation plans and their implementation, formative and summative evaluation. We lead the evaluation work package, or a task inside it.
How people experience, use and trust AI systems, and what changes in their tasks, skills and well-being. Instruments, baseline, pilot and follow-up designs, experience sampling and mixed methods, so the human outcomes of a pilot are measured instead of assumed.
A measurement framework and an objective test of AI literacy, built to compare across languages and countries, with two studies in the field now. Methods to assess skills and skill gaps that stay comparable across sectors and countries.
Evidence on whether an AI system, or an assessment built on it, treats groups equally: subgroup performance and error gaps with confidence intervals, intersectional checks, and measurement invariance, structured for AI Act technical documentation.
How people learn at work and what follows, drawing on our work with the European Company Survey and Cedefop's workplace learning survey. Selection and assessment-centre support for pilot organisations.
From construct to fielded instrument, with multilingual versions tested for comparability, and audits of instruments a project already uses.
Three steps, agreed with the consortium before any data is touched.
1. Define
Which groups, which decisions or outputs, and what counts as unfair in this use.
2. Test
Subgroup performance, error and calibration gaps with uncertainty for models and decision tools, including combinations of groups; invariance and differential item functioning for assessments and scales.
3. Report
Findings with their uncertainty, remediation options, and an evidence file written to the structure that AI Act conformity documentation expects.
Public worked example: Ranks without resolution, a measurement audit of a multilingual language-model benchmark, with data, code and results open. Open the data and code
Self-reported against measured AI literacy, in the same people. In the field in 2026 at the Center for Psychology.
Comparability of an objective AI literacy test across languages and countries. In the field in 2026; schools, employers and research groups can join as fielding partners.
Dr Miloš Kankaraš, co-founder, has been principal researcher on more than ten large international studies. Ana Kankaraš is co-founder and director.
Member of the Psychometric Society, the International Test Commission, IACCP and ACM SIGCHI.
A coordinator and consortium preparing a Cluster 4 proposal on human-AI collaboration at work, for instance HORIZON-CL4-2027-02-DIGITAL-EMERGING-52 (first-stage deadline 2 February 2027), or the Apply AI topics that need adoption and human-factors evaluation.
We take the evaluation and measurement work package, or the fairness-evidence task, as a partner from an Associated Country. On request we draft a one-page work-package outline within days, tailored to your topic, and we can explore pilot or survey sites in Montenegro and the Western Balkans.
