# Vignette 16: Zanzibar's JozaniAI
When annual surveys revealed that only around 15% of tourists returned to Zanzibar after first visit, UNDP Accelerator Lab Tanzania partnered with the Zanzibar Commission for Tourism (ZCT) and the SDG AI Lab to explore a new approach. "We wanted to find ways to build resilience in the sector and competitiveness," explains Tabea Rambuni, Head of Experimentation. For an island where tourism contributes over 50% of the economy and serves as the largest employer, improving visitor experiences was critical to sustainable development.
Instead of relying on periodic surveys alone, the team experimented with analyzing real-time online feedback from platforms like TripAdvisor, Booking.com, and AirBnB using an innovative method known as Aspect-Based Sentiment Analysis (ABSA).[\[1\]](#endnote-1)
The work began in December 2022, with early scoping and co-design sessions. By mid-2023, the team had developed and tested a prototype of the AI model, later named JozaniAI[\[2\]](#endnote-2) (see Figure 31), in collaboration with the UNDP’s SDG AI Lab in Istanbul. The platform categorized tourist sentiments around specific aspects such as hotel cleanliness, service quality, Wi-Fi, and transportation, assigning positive, negative, or neutral scores to each. This allowed policymakers and service providers to go beyond general impressions and act on highly targeted feedback.
Key insights quickly emerged during experimentation. Visitors praised the natural beauty and friendliness of locals but raised concerns about internet access, inconsistent service quality, and transport reliability. In response, several businesses validated the responses during stakeholder engagements. The Zanzibar Commission for Tourism sought to use the data to adjust training priorities and inform promotional strategies. At such a short period of time, the initiative has accelerated service improvement cycles and shifted the culture toward evidence-based decision-making by moving from static data to dynamic intelligence.
The co-creation process was central to the platform’s success. Through hackathons, policy lab engagements, and continuous feedback loops, over 80 stakeholders, from government agencies to youth-led tech hubs, shaped the platform's design and application. This participatory process built local ownership and ensured that the dashboard’s visualizations and indicators were aligned with actual operational needs.
While initially built for tourism, the platform’s design as a digital public good, with open-source architecture and API integration, attracted attention from other sectors. Transport regulators, municipal planners, and health sector innovators expressed interest in adapting the model for citizen feedback. In early 2024, Malawi’s UNDP Accelerator Lab initiated its own pilot using the same ABSA methodology to explore tourist experiences along Lake Malawi.
In Zanzibar, the model also inspired spin-off experiments. The platform was used to identify under-visited attractions and develop targeted interventions, including the launch of the Kizimkazi Hiking Loop and a mobile app built in collaboration with the State University of Zanzibar. These efforts combined AI, community mapping, and local tourism promotion to unlock value in lesser-known destinations.
Over time, the data culture has deepened. Tourism actors now routinely consult the dashboard for decision-making, and youth innovators have gained new skills in AI and data storytelling. What began as a diagnostic tool has evolved into a strategic asset, accelerating innovation and policy agility in Zanzibar’s tourism sector.
{% hint style="info" %}
#### **Key takeaways:**
* **Begin with needs, not technology:** Identify specific challenges, like low tourist return rates, and understand what stakeholders need to improve before designing technological solutions
* **Build for sustainability:** Follow digital public goods principles to create solutions that can be adapted and reused by others
* **Focus on actionable insights:** Transform complex data into clear visualizations that enable evidence-based decision making
* **Create feedback loops:** Regular sprints and stakeholder sessions help capture evolving needs and ensure continuous improvement
* **Plan for skill development:** Ensure users have the capability to effectively utilize new data tools and platforms.
{% endhint %}
***
## Notes
1. For more on the initial stage of this initiative, see the blog by Peter Nyanda (2024). [↑](#endnote-ref-1)
2. [↑](#endnote-ref-2)
---
# Agent Instructions: Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter:
```
GET https://bas-leurss-personal-organization.gitbook.io/undp-accelerator-labs/doing-r-and-d/5.-r-and-d-practices/analyzing-data-and-generating-insights/vignette-16-zanzibars-jozaniai.md?ask=
```
The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.