Top 26%
in this catalogue (246 scored practices)
Scores cluster high, so position within the catalogue is often more telling than the number alone.
Evidence of impact / measured public value2/3
The pilot logged a detailed funnel of results (5,163 called, 2,937 contacted, 646 flagged, 143 completed home visits) but these are engagement/output counts self-reported by the city, with no independent evaluation of actual loneliness-reduction outcomes.
Transparency, fairness & accountability2/3
Madrid's council published the pilot's process and numeric funnel in an official press note and kept the AI as a 'first filter' feeding human social workers, but no algorithm, vendor, or data-protection documentation was found.
Transferability / demonstrated replication1/3
Paloma 2.0 is the same city reusing and expanding its own tool rather than a documented case of another municipality or country adopting it.
Scalability beyond pilot3/3
The city council has already approved and funded a roughly 65-fold expansion in call volume (5,163 to ~337,000) and widened eligibility from over-75s to all over-65s, set to launch September 2026.
Governance, capability & sustainability2/3
The programme is run by a named, budgeted municipal department (Delegación de Políticas Sociales, Familia e Igualdad / Madrid Salud) that carried it from a two-week pilot into an institutionalised, funded service.
Evidoria. Paloma — El sistema de IA de Madrid para detectar la soledad no deseada en personas mayores. Persistent ID: e093e782-0b8e-47fb-9d50-397e944ad30f.