A Kenyan-built artificial intelligence tool is helping people with visual impairments identify obstacles as they move through Nairobi, using a combination of AI object detection, audio guidance, and GPS navigation. The platform, called Sightra, is a web application developed by Kenyan founders Ruth Nzuki and Brian Gillo, designed around the realities of African infrastructure rather than imported from elsewhere.
As users walk, the system detects obstacles ahead and converts that information into speech delivered through an earpiece. GPS technology guides them from a starting point toward a destination. The tool does not replace a mobility cane; it adds information about obstacles further ahead than a cane can detect.
Testing in real conditions
Lawyer Julius Mbura, who has a visual impairment, has been testing the system while moving between his car park and office in Nairobi. He said the real-time audio descriptions help him identify obstacles including parked vehicles and trees. Mbura described the technology as promising while also noting that improvements are still needed in its accuracy and descriptions.
That willingness to identify shortcomings matters because assistive technologies need to be reliable enough to support real-world movement. The Nairobi environment creates particular technical challenges: open drains, missing kerbs, broken surfaces, crowded spaces, parked vehicles, and sudden changes in the physical environment.
The Sightra team says its AI models have been trained and tested around Kenyan infrastructure instead of relying solely on environments with smoother roads and more predictable pedestrian routes. That local focus could prove significant, as technologies developed elsewhere can fail when they encounter environments different from those used during training.
Access and affordability
Kenya's 2022 Demographic and Health Survey reported that around 2 per cent of people aged five and above had some difficulty seeing. Developers say Sightra is already being tested through partnerships with schools serving visually impaired students as well as individual users.
The service uses a subscription model costing $100 annually, although the first six months are currently free. Affordability will remain important; an innovation can only expand independence if the people who need it can access the technology and if the system performs reliably in everyday settings.
There is also a wider policy question. Technology can help users navigate inaccessible streets and buildings, but it does not remove the responsibility of governments and developers to make infrastructure more accessible in the first place. Kenya's National Council for Persons with Disabilities has acknowledged that roads, buildings, and other parts of the physical environment remain difficult for people living with disabilities.
Sightra represents the kind of innovation that becomes possible when local developers begin with local problems. African technology does not always need to imitate products built elsewhere. Sometimes its greatest advantage comes from understanding the street, pavement, drain, building, or community that global developers never designed for.
Sightra is still developing, and its creators acknowledge that more work is needed. But that is how useful innovation grows: it begins with a real problem, a locally informed solution, and the belief that technology should expand independence rather than leave people behind.
Why this matters for customer support professionals
For customer support teams, Sightra offers a useful case study in designing services around the actual conditions users face. The same principle applies to support work: tools and scripts built for one environment often fail when applied to another. AI for Customer Support is increasingly about adapting to real user contexts, not just automating standard responses. The Sightra approach - training on local conditions, testing with real users, and acknowledging limitations - mirrors what effective support teams do when they build knowledge bases and troubleshooting flows for specific audiences. Professionals working in user support can apply this lesson directly: start with the problems your users actually experience, not with a generic solution. An AI Learning Path for User Support Specialists can help teams build these skills systematically.
Your membership also unlocks:
