Across Sub-Saharan Africa, community health workers (CHWs) are already saving lives. Artificial intelligence (AI) can help them save far more.

Frontline workers in many countries are doing extraordinary work. But without the right tools, data, and support, even the most dedicated worker cannot be everywhere care is needed. At Living Goods, we believe AI can help close those gaps. It can enable health systems to deliver faster, more targeted, and continuous care without replacing health workers.

Designing for Performance

Governments across the continent have invested heavily in digitizing community health. Mobile apps and electronic devices have largely replaced paper-based tools, allowing frontline workers to deliver faster care, capture more accurate data, and enhance links to the broader health system.

Despite immense progress, too many digital systems remain just that—digital, not intelligent. They store data, but don’t learn from it. They record what happened, but don’t shape what happens next.

The evolution from digital to intelligent health systems powered by AI faces similar risks. Its impact will depend less on how sophisticated the technology is than on how well it is connected to the people, data, infrastructure, and systems around it.

Drawing on a decade of partnering with governments to design digital tools and national electronic health information systems, we at Living Goods know that even the most promising innovations will struggle to deliver consistent results unless they are designed to perform within real-world constraints. And our approach to AI is no different.

Every tool we build is designed to integrate with government systems and be sustained by governments over time. That is why we treat cost as a parameter from the start—not just how much to invest, but where investment will have the greatest impact. We also lead with a clear understanding of the specific actions that drive health outcomes—not just what works, but why. And we focus our AI innovations on the practical barriers that limit performance.

Laying the Foundations

Across Kenya, Uganda, and Burkina Faso, Living Goods is applying AI-powered solutions to some of the most critical bottlenecks in community health.

One opportunity is seeing problems early enough to act. Our predictive model helps frontline teams identify children at risk of missing one or more recommended vaccine doses, leaving them vulnerable to preventable diseases. The model achieved 78–96% accuracy, giving CHWs actionable insights to intervene before missed visits become dropped immunizations.

Another is the scale of supervision. Although supervision is a critical driver of CHW performance, supervisors cannot personally follow up with every health worker or household. In Kenya, our Smart Supervision Assistant combines data-driven task prioritization with AI-powered features. Integrated with the electronic Community Health Information System (eCHIS), the Assistant analyzes frontline data to flag missed care opportunities and translate them into prioritized follow-up tasks.

Teams using the Assistant have reported encouraging improvements: household visitation increased from 72% to an average of 92–94%, while postnatal care coverage rose from approximately 60% to 85%. Supportive supervision also increased, from approximately 70% to between 80% and 90%. We are rapidly advancing the tool with AI-powered features to provide timely, personalized support to CHWs.

We are also exploring how AI can make frontline learning more responsive to individual needs. Living Goods’ digital learning platform in Kenya has already demonstrated strong demand for flexible, offline-enabled learning, achieving 95% course completion, and 85% monthly active usage in a pilot in Busia County. The next step is to make that learning more adaptive by using AI-powered modules to identify and address individual knowledge and skill gaps, rather than providing every health worker with the same training. For community health workers operating in diverse environments, often without immediate access to a supervisor, this could translate into stronger skills, greater confidence, and ultimately better-quality care for the communities they serve.

Taken together, these applications point to a larger opportunity. AI can help anticipate where care will be missed, direct scarce supervisory capacity where it matters most, and give frontline workers more tailored support.

The real opportunity is not to scale any one of these tools individually. It is to connect them into a system that improves performance across the entire pathway of care.

AI Performance That Scales

Living Goods already knows what stronger frontline performance can achieve. Our approach showed a 27% reduction in under-five mortality. AI creates an opportunity to extend that impact further—helping community health systems reach more families, better deploy the limited number of health workers and supervisors, and ultimately deliver better outcomes at a much greater scale.

Doing so means looking beyond individual AI applications to the persistent bottlenecks that limit performance across the system. Two sit at the heart of community health: families don’t seek care until it is too late, while the supervision needed to help CHWs respond effectively remains difficult to fund and sustain at national scale.

To address this dual challenge, Living Goods is developing an integrated system that combines both patient- and provider-facing AI tools. The first is CareLoop, a free, 24/7 triage and follow-up service that families can access through WhatsApp, voice call, or SMS—all in their own language. CareLoop helps assess potential danger signs and determine the appropriate next step, whether that means managing the illness at home or seeking care at a facility. When needed, it can alert the nearest available CHW and track the response, helping ensure that the patient receives the care they need and that follow-up does not stop there.

For supervision, we are adding an AI agentic layer to our Smart Supervision Assistant that will analyze performance data, generate daily priorities and alerts, handle routine follow-up, and escalate cases requiring human support. This will enable limited supervisory capacity to go much further.

The outlook is promising. But ensuring sustainable results at scale requires the underlying health system to be ready to use, govern, finance, and improve AI tools over time.

The Way Forward

Our experience has reinforced a fundamental lesson: AI can accelerate performance, but it cannot substitute for the systems that make performance possible. Community health information systems across Africa vary significantly in reliability, completeness, and governance. Where foundational systems are fragile, even the best AI tools will underperform or fail to gain government trust.

That is why our approach is not to create another collection of disconnected tools. We are exploring how AI can work with and reinforce the system foundations governments have already invested in. Data architecture, interoperability, governance, and the capacity for stewardship are at the core of our AI approach, not an afterthought.

The goal is government-owned performance that holds as systems grow, making better use of every worker, dollar, and piece of information available to them—so that when families need healthcare, it is there. AI can help us get there, but only when designed around the realities of health systems and built to strengthen them over time. Because ultimately, all roads lead to performance. AI included.