What is itHow it worksFAQ
Our TeamOur PartnersOur Allies
In the mediaResourcesOur newsWebinars & EventsCertificationsSecurity Policy
The Stoplight in ResearchKiana's Mission: From Prison to Purpose
Business without povertyEducation StoplightGreen Stoplight
JOIN THE COMMUNITY
English

The Human-Centered Approach in the Time of AI

Jul 8, 2026

Article by Poverty Stoplight Innovation Week

In this session, Julia Corvalán, Global Operations Manager for the Poverty Stoplight, spoke with Eric Kaduru, Senior Technical Director for Strategic Partnerships, Digital Inclusion, and Inclusive Finance at CARE USA.

Drawing on his experience with savings groups and communities in East Africa, Eric reflected on how to incorporate artificial intelligence and digital tools without displacing people from the center. The conversation addressed designing around local needs, the use of WhatsApp and radio, financial inclusion, and the importance of turning data into useful knowledge for communities.


CHANGE BEGINS WITH COMMUNITIES

Eric explained that although he currently works for an international organization, he began his career as a social entrepreneur in Uganda. This experience allowed him to understand the value of combining the resources of large institutions with the knowledge of those who work directly with communities.

“The approach must begin from the bottom up. That is where change truly starts.”

Local organizations understand people’s practices, needs, and limitations. They should therefore participate in program design rather than merely implement decisions made in other contexts.

Collaboration makes it possible to combine funding and scaling capacity with practical experience, trust, and local knowledge. For Eric, this relationship is essential for designing solutions that truly work.


PEOPLE FIRST, TECHNOLOGY SECOND

In response to concerns that artificial intelligence could replace people, Eric maintained that human knowledge remains essential.

AI systems learn from information and experiences produced by people. For a tool to be relevant, it must incorporate local realities and the experience of those who understand the problem.

Digital transformation should not begin by asking what technology can do, but by asking what people need.

In a community where many people are still learning how to use a smartphone, for example, it would make little sense to begin with an advanced AI tool. The first step could be to use AI to prepare simple content on how to use the device and then gradually progress toward other capabilities.

“The needs of the community must be the starting point.”

Technology must adapt to each population’s levels of connectivity, digital literacy, access to devices, and languages.


AI AS A DOUBLE-EDGED TOOL

Eric acknowledged that AI can accelerate tasks that previously required weeks or months. It can now be used to draft proposals, translate materials, analyze information, and produce content more quickly.

However, it can also generate documents and programs that are disconnected from reality. A proposal may be well written but lack the experiences and knowledge of those who have worked directly with communities.

“AI does not replace the experience of having been there.”

For this reason, people with specialized knowledge should learn to use these tools to expand their capabilities rather than completely delegate their judgment.

AI can prepare a draft or identify patterns, but human review remains necessary to ensure that the result is accurate, relevant, and respectful of the context.


TURNING DATA INTO USEFUL KNOWLEDGE

One example shared during the session was related to financial inclusion and the use of data to establish credit profiles.

AI can analyze the digital records of a savings group and produce a score. However, providing only a number does not necessarily strengthen people’s decision-making capacity.

The tool should explain:

• Why maintaining good records is important. • How a score is calculated. • Which actions can improve it. • Which financial services it could unlock. • What responsibilities are involved in accessing credit.

The objective is not only to process data, but to return it in an understandable form so that people can use it.

This reflection is directly connected to the Poverty Stoplight. Data reaches its greatest value when it helps families interpret their situation, recognize progress, and make decisions about their priorities.


FAST, LOCAL, AND ACCESSIBLE CONTENT

Eric highlighted the potential of AI to produce materials adapted to different languages and formats.

In the past, preparing training in a local language could require several months of translation, recording, and production. Today, technology makes it possible to accelerate the creation of audio content, videos, educational materials, and stories for radio or WhatsApp.

However, the content must emerge from real experiences. AI can facilitate production, but the stories need a human and local foundation.

Eric shared an initiative developed in Uganda to support women as they began using digital tools. The team created a radio series about a young woman who purchased a smartphone and began discovering its different uses.

Each episode presented challenges related to family, work, agriculture, or access to services. The stories were built from situations shared by the communities themselves.

For example, the experience of a woman whose husband initially opposed her use of a phone could be transformed into an episode about how technology enabled her to access agricultural information.

AI helped prepare and adapt the scripts quickly. Radio provided an accessible and familiar channel for the population.

“A story that reflects community life can generate conversations within every household.”

The series combined entertainment and education, allowing people to identify with the characters and reflect on situations similar to their own.


WHATSAPP AS A COMMUNITY TOOL

Eric noted that organizations do not always need to create new applications. In many cases, they can use platforms that are already part of everyday life, such as WhatsApp.

This tool makes it possible to:
• Create learning communities.
• Connect facilitators from different territories.
• Share audio, video, and other materials.
• Ask questions and conduct quick surveys.
• Solve problems through peer support.
• Receive information from the field.
• Offer short training sessions.

Eric created a WhatsApp community that began with facilitators from Uganda and later included people from Kenya and other contexts.

When someone faces a challenge, they can consult the group and receive recommendations from people with similar experiences. This generates collective learning and makes it possible to improve training based on real problems.

WhatsApp also makes it easier to obtain information without waiting for a formal field visit. Facilitators can share audio messages, photographs, or videos from communities.

Eric clarified that this information does not replace rigorous monitoring and evaluation processes, but it can help identify difficulties, adjust activities, and document lessons more quickly.

Julia connected this experience with the Poverty Stoplight. Families can update their indicators when they achieve a change without waiting several months for a new assessment. This makes it possible to recognize progress and adjust support more promptly.


MOTIVATION AND LEARNING

Eric also highlighted the value of gamification in digital tools.

An animation, a congratulatory message, or recognition for completing a task may seem unnecessary for an adult. However, these elements increase motivation, particularly among people learning how to use a tool.

A Poverty Stoplight indicator turning green can become a moment of celebration and motivate other members of the group.

Julia mentioned competitions organized by Fundación Paraguaya in which committees work together to help a person improve indicators related to bathrooms, dental health, or other conditions.

Motivation arises both from recognition and from collective support, as well as from seeing that change is possible.


FROM BENEFICIARIES TO ECONOMIC ACTORS

Eric noted that savings groups already move money, issue loans, and manage resources. However, much of this activity remains invisible to financial institutions because it takes place informally.

Digitizing records makes it possible to demonstrate the groups’ capacity for saving, repayment, and organization.

“Technology can make visible an economic system that already exists.”

This changes the relationship with banks. Communities stop being viewed solely as beneficiaries in need of assistance and begin to be recognized as clients and economic actors.

Based on this information, financial institutions can design products suited to the groups’ actual needs.

Technology therefore acts as a bridge between formal financial systems and the economic practices that already exist within communities.


BALANCING BENEFITS AND RISKS

Eric acknowledged the importance of privacy and information security. However, he warned that organizations must also listen to the concrete benefits people find in a tool.

For a woman who wants to communicate with relatives, sell a product, or access agricultural information, WhatsApp may address an immediate need.

Organizations must manage risks without blocking important benefits. This requires:
• Clearly explaining how data will be used.
• Obtaining consent.
• Protecting sensitive information.
• Using tools that people can understand.
• Maintaining alternatives for those without digital access.
• Defining priorities together with communities.

The design process must recognize that not all risks and needs can be addressed at the same time. Decisions should be made with people, not solely within institutions.


THE VOICES MISSING FROM ARTIFICIAL INTELLIGENCE

During the question-and-answer segment, participants noted that large language models still do not sufficiently incorporate the experiences of communities experiencing poverty.

Eric agreed that there is an opportunity to document local stories, solutions, and lessons.

The Poverty Stoplight can contribute because it generates information directly with families about their conditions, priorities, and actions.

Even people without connectivity can contribute knowledge through conversations with facilitators. However, this process must respect privacy, consent, and ownership of information.

Communities must not become merely sources of data for systems built by others. They should participate in determining how their experiences are collected, interpreted, and used.


Get Involved

Interested in attending future virtual and in-person events from the Global Stoplight Network? Subscribe to our monthly newsletter here.

You can learn more about the upcoming Cerrito Forum 2026, the annual international global development event that brings together global stakeholders committed to the elimination of poverty from October 12th - 16th by visiting cerrito.com.py.

Back to all webinars
FOLLOW US
WHAT WE DO
What is it
How it works
FAQ
WHO WE ARE
Our Team
Our Partners
Our Allies
RESOURCES
In the news
Publications
Our blog
Privacy Policy
Terms And Conditions
STAY IN TOUCH

(595) 21 609 277  |  

info@povertystoplight.org

Manuel Blinder 5589
Asunción, Paraguay

POWERED BY

© Copyright 2020 — Fundación Paraguaya

Made with love byPenguin Digital