You May Also Be Interested In
When Innovation Meets Reality: Validating a Diesel Discount App
How we saved $50,000 and months of development by validating a voice-only app for Nigerian truckers
Read case studyDeep insights from designing a voice-first UX for Nigerian truck drivers, and implications for an AI-led future
Voice as Primary Interface Layer
Cultural-First Design
AI Integration Ready
Low literacy, across the developing world, is more than an educational gap - it is an accessibility challenge. My beloved grandmother couldn't share in the lives of her great-grandchildren via smartphones simply because she was illiterate.
In Nigeria, you'll see this manifest daily in how semi-literate workers struggle to use text-heavy smartphone apps like WhatsApp:
Text-based interfaces either exclude them entirely or serve them poorly.
The fundamental innovation in our project was reimagining voice not as a supplementary feature, but as the primary interface layer. This represents a profound shift from traditional UX design:
Instead of text tutorials or help documentation, voice becomes the constant teacher, actively guiding users through each interaction. Every screen starts with just a help button that, when pressed, triggers voice explanation of the screen's purpose and functionality.
Rather than static text prompts, voice provides dynamic, contextual guidance. When users input a phone number, for instance, the voice system actively confirms correct inputs and immediately alerts users to errors, creating a conversational, supportive interaction model.
The system uses voice to actively manage user anxiety and build confidence. In error states, voice explicitly tells users not to panic and assures them that their "teacher" will help resolve any issues.
Instead of presenting all UI elements at once with text labels, voice works in concert with progressive disclosure. UI elements only appear as they're explained by voice, creating a synchronized learning experience.
Traditional UX research methods would have been too costly for developing my MVP. Instead of deploying professional researchers, I recruited an intelligent truck driver as our researcher.
Initially, I doubted whether my semi-literate driver could deliver detailed feedback. But through iterative coaching - emphasising storytelling with rich detail - I discovered he could provide surprisingly detailed insights.
Creating a voice-led UX for users with no prior smartphone experience meant discarding assumptions. Familiar icons, text-based cues, or complex interactions were off the table. Instead, design decisions relied on metaphors from their daily lives.
Design with the baseline assumption that users cannot read or understand visual iconography. The only safe assumption: users can recognize numbers (due to currency handling). Everything else we take for granted is potentially alien.
In Nigeria, storytelling often relies on relatable examples. We consistently leveraged this through:
The emergence of advanced AI technology opens up exciting new possibilities for voice-led interfaces. Here's how AI could transform this space:
Instead of pre-recorded voice prompts, AI could generate contextual guidance in real-time, enabling:
Modern language models could enhance the system's ability to bridge cultural gaps:
Early failures taught us crucial lessons. And most came from the language selection screen.
For example, we used headshots of people in traditional headdress to represent 4 main languages in a grid layout, but this confused users:
We learnt that:
These changes provided multiple identification cues (position, features, headdress) and proved crucial to ease of use.
My core learning from the issues with language selection was that cognitive overload must be made redundant by design.
So I decided that every screen would start off blank with just a help button as the dominant feature. This approach reassures the user that they will always be handheld.
The user journey to this realisation was designed as follows:
A good example of input validation is mobile number entry:
This work suggests mobile apps should evolve toward three user-selectable modes:
AI could make voice-led interfaces more scalable by:
Designing for illiterate users is not just about technology—it's about empathy and creativity. By reframing the role of voice as a core UX element, we can expand digital accessibility and empower underserved communities.
The next wave of innovation in this space will likely come from combining the human-centered design principles we discovered with the powerful capabilities of AI.
Communicating abstract concepts, such as tiered membership benefits, to illiterate or semi-literate users is a challenging task. During our design process, we successfully used a culturally resonant storytelling approach to explain the concept of membership levels to truck drivers.
Our system offered three levels of membership for truck drivers, each with specific benefits. The challenge was to ensure users understood these levels and their associated rewards without relying on text-heavy descriptions or human intervention.
We combined two familiar metaphors—schooling and fishing—to guide users through the concept:
We began with a picture of children in a classroom and explained that the membership system was like a school with multiple levels. Each level represented progress, akin to moving to the next grade in school after passing exams.
To make the benefits at each level tangible, we transitioned to the familiar scenario of fishing:
A picture of a fisherman in a dugout canoe, paddling on open water without a catch, symbolized the starting level.
Next, we showed a fisherman with a modest-sized fish to represent initial benefits.
Finally, we presented a fisherman with a significant catch to represent premium benefits.
The approach proved remarkably successful:
This case study demonstrates how combining cultural metaphors, progressive disclosure, and synchronized voice guidance can make complex concepts accessible to illiterate users. The success of this approach suggests that similar metaphor-based explanations could be effective for other abstract concepts in voice-led interfaces.
How we saved $50,000 and months of development by validating a voice-only app for Nigerian truckers
Read case study