ESP-002

Paloma / Paloma 2.0 – AI Virtual Assistant for Elderly Loneliness Detection

Download PDF
Spain Europe & Central Asia High income Operational Deployment (Limited Rollout) Confirmed

Madrid City Council (Ayuntamiento de Madrid) – Department of Social Policies, Family and Equality (Área de Gobierno de Políticas Sociales, Familia e Igualdad); Directorate General for Elderly and Prevention of Unwanted Loneliness (Dirección General de Mayores y Prevención de la Soledad no Deseada)

At a Glance

What it does Classification — Vulnerability, needs and risk assessment, including predictive analytics
Who runs it Madrid City Council (Ayuntamiento de Madrid) – Department of Social Policies, Family and Equality (Área de Gobierno de Políticas Sociales, Familia e Igualdad); Directorate General for Elderly and Prevention of Unwanted Loneliness (Dirección General de Mayores y Prevención de la Soledad no Deseada)
Programme Paloma / Paloma 2.0 – AI Virtual Assistant for Elderly Loneliness Detection
Confidence Confirmed
Deployment Status Operational Deployment (Limited Rollout)
Key Risks Not assessed
Key Outcomes Pilot (Nov-Dec 2023): 5,163 persons contacted; 2,937 (57%) answered; 2,071 (71% of respondents) completed conversation; 1 in 3 reported feeling lonely; 646 (35%) requested follow-up; 378 contacted for home visits, 236 accepted, 143 home visits completed; 84% of visits confirmed loneliness situation; 44 high-risk cases identified for intensive Social Services intervention.
Source Quality 3 sources — Report (multilateral / development partner), News article / media, Government website / press release

Paloma is an AI-enabled virtual assistant developed by Madrid City Council (Ayuntamiento de Madrid) as part of the municipality's strategy to combat unwanted loneliness among elderly residents. The system uses natural language processing (NLP) technology to conduct automated telephone calls that simulate human conversation, enabling proactive outreach to elderly citizens who may be at risk of social isolation.

The initial pilot project ran between 28 November and 8 December 2023. During this period, the system placed calls to a sample of 5,163 persons aged 75 and over who were identified as living alone based on municipal registry data. The sample comprised 70% women and 30% men, with an average age of 81 years. Of those called, 2,937 persons (57%) answered, and 2,071 (71% of respondents) completed the full conversation. In total, the virtual assistant logged 113 hours of dialogue with elderly residents, the majority residing in the districts of Moncloa-Aravaca, Villaverde, and Villa de Vallecas.

The system administered five structured questions designed to assess loneliness risk. The first question asked whether the person often felt lonely, to which 696 out of 2,104 respondents (33%) answered affirmatively — 31% of women and 34% of men. The second question asked whether they had family or friends to talk to when worried; 93% of 1,935 respondents confirmed they did, though notably 85% of those who reported feeling lonely also had such support networks. The third question asked whether they had someone to turn to in case of need, with 88% of 1,872 respondents answering yes; however, 78% of those who felt lonely also had someone available. The fourth question asked whether they went out at least once a week, with 93% of 1,847 respondents saying yes, though 33% of these still reported feelings of loneliness. Among those who did not leave home weekly, 63% received regular visits but 46% still felt lonely, while 61% of those receiving no visits felt lonely. The fifth and final question asked whether they would like a municipal technician to contact them for a situation assessment; 646 persons (35%) accepted this offer.

Following the pilot's initial AI-based detection phase, the municipality deployed professional follow-up interventions. Of the 646 persons who accepted follow-up, 602 who were not considered high-risk received individualized attention through the municipal programme 'Acompañamiento a la integración social de personas mayores que se sienten solas' (Accompaniment for Social Integration of Elderly Persons Who Feel Lonely). Of these, 378 had been contacted and 236 had accepted home visits, of which 143 had been completed at the time of the February 2024 press release. Additionally, 44 persons identified as high-risk were referred for comprehensive assessment by Social Services. Overall, in 84% of home visits conducted, the loneliness situation was confirmed.

The results were presented in February 2024 by José Fernández, the delegate for Social Policies, Family and Equality, as part of the broader Municipal Strategy Against Unwanted Loneliness (Estrategia municipal para combatir la soledad no deseada en las personas mayores). The municipality noted that AI enabled a substantial change in processing complex information in real time while achieving broad reach and adherence.

In November 2025, a scaled-up version called Paloma 2.0 was announced. Paloma 2.0 expands the outreach to 100,000 planned telephone calls and is integrated with the Madrid City Council social history database (historia social del Ayuntamiento de Madrid), enabling more personalized, empathetic, and needs-adapted care. The updated system adapts the frequency of follow-up calls based on the degree of loneliness detected and automatically refers severe cases to specialized services. Paloma 2.0 was a finalist for the European Social Services Awards, presented at a ceremony held in Madrid in November 2025.

Classifications follow the DCI AI Hub Taxonomy. Hover over field labels for definitions.

Social Protection Functions

Implementation/delivery chain
Outreach/communications/sensitisation primaryAssessment of needs/conditions + enrolment
SP Pillar (Primary) The social protection branch: social assistance, social insurance, or labour market programmes. Social assistance
Programme Name Paloma / Paloma 2.0 – AI Virtual Assistant for Elderly Loneliness Detection
Programme Type The type of social protection programme, classified under social assistance, social insurance, or labour market programmes. View in glossary Other
System Level Where in the social protection system the AI is applied: policy level, programme design, or implementation/delivery chain. View in glossary Implementation/delivery chain
Programme Description Municipal AI-enabled outreach programme for proactive identification and support of elderly persons (75+) living alone who may be experiencing unwanted loneliness. The programme uses an AI virtual assistant (Paloma/Paloma 2.0) to conduct automated telephone calls, assess loneliness risk through structured questions, and refer identified individuals to appropriate municipal social services including home visits, social integration programmes, and specialised Social Services assessment for high-risk cases.
Implementation Type How the AI output is produced: Classical ML, Deep learning, Foundation model, or Hybrid. Affects validation, compute requirements, and governance profile. View in glossary Deep learning
Lifecycle Stage Current stage in the AI lifecycle, from problem identification through to monitoring, maintenance and decommissioning. View in glossary Integration and Deployment
Model Provenance Origin of the AI model: developed in-house, adapted from open-source, commercial/proprietary, or accessed via third-party API. View in glossary Not documented
Compute Environment Where the AI system runs: on-premise, government cloud, commercial cloud, or edge/device. View in glossary Not documented
Sovereignty Quadrant Classification of data and compute sovereignty: I (Sovereign), II (Federated/Hybrid), III (Cloud with safeguards), or IV (Shared Innovation Zone). View in glossary Not assessed
Data Residency Where the data used by the AI system is stored: domestic, regional, or international. View in glossary Not documented
Cross-Border Transfer Whether data crosses national borders, and if so, whether documented safeguards are in place. View in glossary Not documented
Decision Criticality The rights impact of the decision the AI supports. High criticality requires HITL oversight; moderate requires HOTL; low may operate HOOTL. View in glossary Moderate
Human Oversight Type Level of human involvement: Human-in-the-Loop (active review), Human-on-the-Loop (monitoring), or Human-out-of-the-Loop (periodic audit). View in glossary HITL
Development Process Whether the AI system was developed fully in-house, through a mix of in-house and third-party, or fully by an external provider. View in glossary Not documented
Highest Risk Category The most significant structural risk source identified: data, model, operational, governance, or market/sovereignty risks. View in glossary Not assessed
Risk Assessment Status Whether a formal risk assessment, informal assessment, or independent audit has been conducted for this system. Not assessed

Impact Dimensions

Equality, non-discrimination, fairness and inclusion
Systemic and societal
  • Human oversight protocol
CategorySensitivityCross-System LinkageAvailabilityKey Constraints
Administrative data from other sectorsPersonalLinks data across multiple systemsCurrently available and usedLimited to persons registered in the municipal registry (padrón municipal) as aged 75+ and living alone; depends on accuracy and currency of municipal registry records
Beneficiary registries and MISSensitiveLinks data across multiple systemsCurrently available and usedAccess restricted to municipal social services staff; integration enables personalised care but raises data protection considerations
Unstructured and text-based contentSensitiveLinks data across multiple systemsCurrently available and usedConversational data collected via AI telephone calls with structured questions; quality depends on respondent comprehension and willingness to engage; 113 hours of dialogue recorded in pilot

Lara-Montero, A. (2024) 'The Transformation Potential of AI on Social Services', European Social Network. Available at: https://www.esn-eu.org/news/transformation-potential-ai-social-services (Accessed: 22 March 2026).

View source Report (multilateral / development partner)

Luengo, A. (2025) 'El sistema de Inteligencia Artificial de Madrid que ayuda a los más mayores con atención personalizada', El Debate, 22 November. Available at: https://www.eldebate.com/espana/madrid/20251122/sistema-inteligencia-artificial-madrid-ayuda-mayores-atencion-personalizada_357740.html (Accessed: 22 March 2026).

View source News article / media

Ayuntamiento de Madrid (2024) 'Madrid inicia la atención social a más de 600 mayores en soledad no deseada detectados mediante Inteligencia Artificial', Diario del Ayuntamiento de Madrid, 2 February. Available at: https://diario.madrid.es/blog/notas-de-prensa/madrid-inicia-la-atencion-social-a-mas-de-600-mayores-en-soledad-no-deseada-detectados-mediante-inteligencia-artificial/ (Accessed: 22 March 2026).

View source Government website / press release
Deployment Status How far the system has progressed into real-world operational use, from concept/exploration through to scaled and institutionalised. View in glossary Operational Deployment (Limited Rollout)
Year Initiated The year the AI system was first initiated or development began. 2023
Scale / Coverage The scale and geographic or population coverage of the deployment. Pilot: 5,163 persons contacted (Nov-Dec 2023); Paloma 2.0: 100,000 planned calls (announced Nov 2025)
Funding Source The source(s) of funding for the AI system development and deployment. Unknown
Technical Partners External technology vendors, academic partners, or development partners involved. Not specified in sources; developed in collaboration with Madrid City Council technology partners
Outcomes / Results Pilot (Nov-Dec 2023): 5,163 persons contacted; 2,937 (57%) answered; 2,071 (71% of respondents) completed conversation; 1 in 3 reported feeling lonely; 646 (35%) requested follow-up; 378 contacted for home visits, 236 accepted, 143 home visits completed; 84% of visits confirmed loneliness situation; 44 high-risk cases identified for intensive Social Services intervention. Scale-up (Paloma 2.0): Expansion to 100,000 planned calls; integration with municipal activity/service information; automatic referral for severe cases; frequency of follow-up adapted to detected loneliness level

How to Cite

DCI AI Hub (2026). 'Paloma / Paloma 2.0 – AI Virtual Assistant for Elderly Loneliness Detection', AI Hub AI Tracker, case ESP-002. Digital Convergence Initiative. Available at: https://socialprotectionai.org/use-case/ESP-002 [Accessed: 1 April 2026].

Change History

Updated 1 Apr 2026, 08:11
by system (system)
Updated 31 Mar 2026, 06:35
by system (system)
Created 30 Mar 2026, 08:39
by v2-import (import)