Bolivia's Ministry of Education launched the AI chatbot AFINIDATA nationwide with UNICEF in Aug 2025, giving rural caregivers personalised early-childhood guidance via WhatsApp; a peer-reviewed Peru trial of the same platform found a real 0.11 SD child-development gain.
1,458 mothers and fathers
Caregivers using AFINIDATA via El Alto's DIPI platform (2025)
1,900 children
Children covered by El Alto DIPI usage (2025)
0.11 SD
Child-development effect size, AI chatbot arm (Peru trial, not Bolivia) (child age 2.5 years)
0.17 SD
Child-development effect size, human home-visiting arm (Peru trial, not Bolivia) (child age 2.5 years)
Details
Maturity
Scaling
Promoter
Ministerio de Educacion, Deportes y Culturas (Bolivia); UNICEF Bolivia; Afinidata
Period
Aug 2025-ongoing (Bolivia national launch); RCT evidence from Peru deployment 2021-2025
Keywords
early childhood development, AI chatbot, parenting support, WhatsApp
Context
AFINIDATA is a WhatsApp/Facebook Messenger chatbot, originally developed in Guatemala in 2017, that uses a machine-learning recommender to personalise early-stimulation and health/nutrition guidance for caregivers based on each child's age and developmental stage. In August 2025, Bolivia's Ministry of Education, Sports and Cultures launched AFINIDATA nationwide with UNICEF Bolivia as the AI component of its 'Aprendiendo en Familia' (Learning as a Family) initial-education programme, targeting families across the country, especially in rural areas, with companion guidance booklets for teachers and families.
Activities
At the municipal level, the El Alto city government made AFINIDATA available free of charge over WhatsApp through its 'DIPI' early-childhood platform, part of the wider 'Pan Manitos' programme for children under four.
Results
During 2025, El Alto reported that 1,458 mothers and fathers responsible for 1,900 children used AFINIDATA through the municipal DIPI platform. Separately, a peer-reviewed cluster-randomized controlled trial of the same underlying platform in rural Cajamarca, Peru (2,461 caregiver-child dyads across 164 clusters) found that both an AI-chatbot arm and a parallel human home-visiting arm improved child development at 2.5 years of age, with standardized effect sizes of 0.11 and 0.17 respectively — the digital intervention achieved a smaller effect than home visiting but at roughly one-fifteenth of the cost.
Conclusions
Bolivia's national rollout is real and operating, but as of this entry only city-level (not national) usage figures have been published and no causal evaluation of the Bolivia deployment itself exists yet; the Peru RCT provides strong evidence for the underlying model but cannot be assumed to transfer automatically to the Bolivian context and delivery.
Implementation
Indicative cost
Low (< €50k) — No Bolivia-specific budget has been published. The comparable Peru evaluation found the digital/chatbot arm cost approximately one-fifteenth of a parallel human home-visiting arm, consistent with a low-cost delivery model, but this is a cost ratio from a different country's trial, not a confirmed Bolivia budget.
Time to results
Short (< 1 year) — Nationally launched in August 2025; as of this entry (less than one year later) only municipal-level (El Alto) usage data for 2025 has been published.
Staffing & skills
Bolivia's Ministry of Education, Sports and Cultures (national programme lead), UNICEF Bolivia (implementation partner), Afinidata (platform developer/technical partner), El Alto municipal government / DIPI programme (local delivery channel)
Conditions for success
Delivery over an already-ubiquitous channel (WhatsApp) removing the need for a dedicated app or new hardware
Companion print guidance booklets for teachers and families supporting the digital tool
Integration with an existing municipal early-childhood programme (Pan Manitos/DIPI) rather than building a parallel structure
Common failure modes
No Bolivia-specific outcome evaluation exists yet — the programme is less than a year old and only one municipality has published usage figures
Strong RCT evidence exists for the platform in a different country (Peru) and should not be assumed to transfer automatically to Bolivia's rural contexts and delivery model
Where it fits
Governance type
national government programme delivered via municipal/local government and international-agency partnership
Scale
national (with only municipal-level data published so far)
Income level
lower-middle-income
Data sources
Where this practice's information was retrieved from, and when.