evidoria

← Back to browse

Good practice Imported

DIVA — Mauritius' Government-Wide AI Chatbot for Public Service Access

Mauritius · Port Louis · See the Mauritius profile · See the Port Louis profile

Evidence: Descriptive / self-reported Top 67% 53/100 · Ask Evidence Copilot about this practice

Mauritius' Central Informatics Bureau runs DIVA, an AI chatbot answering citizen queries for eight ministries via web, WhatsApp and app in English, French and Creole; official reports cite 500+ conversations a month but no independent evaluation of outcomes exists yet.

500+ conversations/month
DIVA conversations handled per month (self-reported)
DIVA — Mauritius' Government-Wide AI Chatbot for Public Service Access

Details

Maturity
Scaling
Promoter
Central Informatics Bureau, Ministry of Information Technology, Communication and Innovation (Mauritius)
Period
2023–2026
Keywords
digital government, citizen engagement, chatbots, e-government, natural language processing

Context

Mauritius' Central Informatics Bureau, under the Ministry of Information Technology, Communication and Innovation, operates DIVA (Digital Interactive Virtual Assistant), a government-wide AI chatbot that has been deployed since 2023 across eight ministries and departments to answer citizen queries and direct users to services.

Objectives

DIVA aims to give citizens a single, multilingual (English, French, Mauritian Creole) point of access to government services via the government website, WhatsApp, a mobile app and a humanoid kiosk interface, reducing reliance on phone calls and in-person office visits. The government has stated DIVA will become the default AI interface for every future government mobile service, including an upcoming chatbot layer for the National Electronic Licensing System (NELS).

Activities

The system uses natural-language-processing and machine-learning components hosted on Amazon Web Services, with its knowledge base updated by scraping government websites.

Results

The platform's own operator reports DIVA handling over 500 conversations per month, replacing some phone calls and physical office visits. No independent evaluation of cost savings, resolution accuracy or citizen satisfaction has been published. In April 2026, Mauritius launched a National AI Strategy and FAIR Guidelines (Fairness, Accountability, Inclusiveness and Integrity, Responsibility), signalling that formal governance and evaluation structures for tools like DIVA are still being built out.

Conclusions

The rubric evaluation (overall score 53.3/100) rates evidence of impact and transparency as weak - usage figures are self-reported and no bias testing or performance disclosure specific to DIVA has been published - while noting the system is already live across eight ministries on a portable, commodity-cloud architecture, operated by a dedicated public body now covered by Mauritius' national AI ethics framework.

Implementation

Indicative cost
Medium (€50k–€500k) — No published cost figures; hosted on commodity cloud infrastructure (AWS) across web, WhatsApp, mobile app and a humanoid kiosk interface - conservative estimate pending curator review.
Time to results
Medium (1–3 years) — Deployed since 2023 across eight ministries/departments; expansion to all future government mobile services (including NELS licensing) and a formal national AI governance framework announced April 2026.
Staffing & skills
Central Informatics Bureau (Ministry of Information Technology, Communication and Innovation, Mauritius)

Conditions for success

  • A dedicated public body (Central Informatics Bureau) operating the platform
  • Multi-channel deployment (web, WhatsApp, mobile app, kiosk) to meet citizens where they are
  • Commodity cloud NLP/ML architecture (AWS) that is well-precedented and portable
  • A National AI Strategy and FAIR Guidelines (April 2026) establishing a governance framework

Common failure modes

  • No published model risk assessment, bias testing, or performance disclosure specific to DIVA
  • No independent evaluation of cost savings, resolution accuracy or citizen satisfaction

Where it fits

Governance type
national government agency
Scale
national (8 ministries, expanding)
Income level
upper-middle-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful