For EU institutions, funders, and research consortia
A partner that already holds the compliance file.
A deployed platform, real data, a documented methodology, and the ethics annex mostly written. This page is here to be evaluated, not to sell.
Positioning, stated plainly
This is public infrastructure that happens to be sold commercially.
Municipalities and employers pay for it, which is what keeps it maintained rather than abandoned at the end of a grant cycle. Nothing in the architecture requires a partner to keep paying us to keep using what they built.
What we bring in
Four things, and the fourth is usually the missing one.
Replication
Three countries the method transfers to, and what breaks in each.
A replication claim without its failure modes is a hope, not a claim.
| Country | Why it transfers | What costs more than expected |
|---|---|---|
| France | Comparable municipal competence over integration and social services, and an established sub-municipal statistical geography in the IRIS units, so the neighbourhood unit of analysis transfers without redefinition. | Employment data sits with a national operator rather than the commune, so the workforce half of the spine needs a national data agreement that the civic half does not. |
| Italy | Municipal registry data at a resolution the method can use, and regional integration programmes already reporting against AMIF indicators, so the reporting layer maps onto an existing obligation. | Registry quality varies sharply between comuni, which moves more work into the diagnostic phase and makes the baseline slower to fix. |
| Portugal | A national integration programme with municipal delivery and an existing indicator framework, plus a language model landscape close enough to the Spanish and Catalan work to reuse most of it. | Smaller municipalities mean smaller cell sizes, and the minimum cell size rule suppresses more of the output. The unit of analysis may have to move up a level. |
This is an assessment, not a result. No deployment outside Catalonia exists yet, and the frictions above are the ones we expect rather than ones we have met.
Academic review
Reviewed before it runs, re-reviewed on a cycle, published back.
The third move is the one a consortium can use: a deployment that returns findings and failure records to a research group is a work package rather than a courtesy, and it is the reason the method is written to be read by someone who did not build it.
Reviewed before it runs
The method goes out for review before it touches anyone’s data.
Index design, baseline construction, and simulation parameters are written as documents and reviewed by a research group that works on that subject. The method is separable from the software on purpose, so it can be evaluated by people who did not build either. An output produced by a method that has not been through this says so on its face.Re-reviewed on a cycle
Annually, and whenever the method changes materially.
Not per client and not per delivery. An engagement runs a version of a method that was already reviewed, and anything an engagement adds goes into the next cycle rather than holding up the work. The version that produced a given output is recorded on it.What we learn goes back
Findings, method changes, and failures return to the research group.
A review is worth something to a research group only if the deployment feeds it. Anonymised findings and the record of what did not work go back, for further research and for the policy questions that research exists to answer. This is the half of the arrangement that makes it a collaboration rather than a favour.
Compliance architecture
The ethics annex, mostly written.
These six statements are architectural rather than procedural: each one is enforced at a boundary in the system rather than promised in a policy document, which is the distinction an ethics reviewer is looking for.
Anonymisation at ingestion
No individual-level data reaches the AI layer.
EU data residency
EU infrastructure, Catalan residency for Catalan deployments.
Per-module EU AI Act classification
Classified per module, before deployment, because they differ.
A human decides
The platform ranks and explains. A person decides.
DPA before ingestion
Signed, and your DPO briefed, before a record moves.
Exportable audit trail
Every output keeps what produced it, and exports.
Published work
What we have put our name to in public.
Each one states its finding, shows its evidence, and names its limitation in the same document.
- Field brief
The six-week window, and where the number comes from
The interval between a neighbourhood signal appearing in administrative data and the point where a response still changes the outcome is about six weeks. Most municipalities are working eighteen months behind it.
- Data
One in four: what the 25.1% figure does and does not tell you
More than a quarter of Catalonia’s residents were born outside Spain. The figure is accurate, widely quoted, and almost useless for planning a service, because the thing it averages away is the thing you need.
- Government
What your Data Protection Officer asks first
Six questions decide whether a municipal AI pilot happens. They are asked in the same order every time, and a vendor who cannot answer the first two rarely gets to the third.
A named contact, not a form.
Write with the call identifier and the deadline. You will get a reply from someone who can commit to a work package, and if the fit is not there you will be told so in the same message.