Unlocking the potential of AI for social health insurance

GIZ, ADB, and Amref Health Africa have joined forces to develop an e-Learning course on integrating AI into the management of social health insurance programmes. A Learning Exchange approach is helping to ensure the course is tailored to real world challenges in Africa and Asia.

 

As countries scale up national health insurance programmes in pursuit of Universal Health Coverage (UHC), governments are increasingly relying on digital systems to manage the fast-growing volumes of data and financial transactions. While traditional digital information management systems can help to collect and analyse data, the shift to Artificial Intelligence (AI) represents a transformative leap for the insurance industry with its potential to automate routine insurance tasks, such as processing claims and preventing fraud.

‘The way forward for us is with AI’, says Rashid Amour, Manager of the Software Development Department at Tanzania’s National Health Insurance Fund (NHIF). He continues, ‘we need AI because we have a very large number of insurance claims to manage on a daily basis.’

The road to AI adoption, however, is rarely straightforward. Health financing programmes, such as health insurance, often face deep-rooted challenges – not only with AI but with digital transformation more broadly. Fragmented databases and continued reliance on paper-based health records pose significant barriers. And these challenges aren’t limited to resource-constrained settings like Tanzania; better-resourced countries frequently lack the expertise to begin effectively planning for AI, as well as the legal and policy frameworks to smooth its introduction.

Mr. Amour was speaking at a capstone workshop in Nairobi in March 2025 – the final stage of an innovative Learning Exchange that brought together health insurance and digital health experts from Cambodia, Ghana, Kenya, Nepal, Nigeria, and Tanzania. This Learning Exchange, and a second Exchange starting in August 2025, are joint initiatives of Amref Health Africa (Amref) and the Global Alliances for Social Protection (GASP) programme of the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). Together they are informing the development of an e-Learning course on AI for Health Insurance – Claims Management & Fraud Control.
 

 

Building AI capacities in partner countries

Germany’s Federal Ministry for Economic Cooperation and Development (BMZ) places a strong emphasis on building digital capacities in partner countries. Since launching the Digital Strategy in 2023, BMZ – and the German government as a whole – has significantly increased support for AI capacity development as a means to help countries fast-track progress towards achieving the Sustainable Development Goals (SDGs).

The GASP programme supports the development and introduction of digital public goods for social protection and health financing, including the openIMIS software. The programme also advances standard-setting to enhance interoperability between social protection systems, strengthens national leadership and ownership of social protection programmes, and promotes bi- and multilateral collaboration among social protection actors.

Together with Amref, GASP is supporting the development of the e-Learning course on AI for Health Insurance, which emerged in response to mounting interest in AI among social protection policy makers and practitioners. The course is designed to strengthen the capacities and skills of professionals working in social health insurance, equipping them to identify potential entry points for AI and to engage with AI in a practical, ethical, and informed way.

What can AI do for health insurance?

Access to good quality, acceptable and affordable healthcare remains out of reach for many people around the world. Governments in low- and lower-middle income countries see AI as a strategic tool to support the scale-up of health insurance, increase efficiencies, and improve equity through better targeting, particularly in the context of limited resources. Growing interest in integrating AI into public sector social protection and social health insurance has highlighted the urgent need to develop AI-relevant skills and capacities, and strengthen enabling environments for AI implementation.

While AI is widely used in private health insurance, public sector health insurance programmes are at an earlier stage in their AI journeys. A number of large national social health insurance agencies, such as Indonesia’s Health Care and Social Security Agency (BPJS-K), and the Philippine Health Insurance Corporation (PhilHealth), have started to experiment with and use machine learning to speed up claims processing and to identify and prevent fraud. National insurance programmes in Cambodia and Nepal are also in the process of implementing AI modules as part of the openIMIS information management software, both of which are scheduled to go live later this year.

Valuable lessons are consequently emerging – not only on how AI can enhance insurance programme management, but also on the policy and governance frameworks needed to guide its responsible use. Key issues include data quality and standards, privacy, ethics and other foundational issues.

Learning grounded in real-world experience

The Learning Exchange approach focuses on real-world problem solving, enabling participants from diverse countries and contexts to come together and to share experience-based insights, challenges and practical solutions. AI expertise was provided by two facilitators, Simona Dobre – Data Scientist and AI Specialist, and Tim Ohlenburg – Consultant in Economics, Data Science, and Machine Learning.

While few of the participants of the first Learning Exchange had much direct experience of working with AI, they contributed valuable experience in health insurance policy and implementation, and the digital health ecosystems of their respective countries. The Learning Exchange takes as its starting point the real challenges participants face in their work. As Simona Dobre explains,

I like to start from the everyday challenges participants face in managing health insurance systems. From there, we explore which processes could be automated using AI – the AI models and practices that could help address these practical concerns.

 

A second Learning Exchange (Aug-Oct 2025) is engaging senior health insurance managers and IT executives from six additional countries – bringing the number of participating countries to twelve. The Asian Development Bank (ADB) is joining GIZ and Amref in hosting the second Exchange, building on their recent collaboration during the Inclusive, Sustainable, Prosperous and Resilient (INSPIRE) Health Systems in Asia and the Pacific Health Forum held in Manila.

Through the regional technical assistance project Using Digital Technology to Improve National Health Financing in Asia and the Pacific, ADB is co-designing the learning agenda, providing expert participation for the in-person workshop, and will support the wider distribution and marketing of the e-learning course.

‘ADB is committed to assisting its member countries expand the use of AI in health financing,’ says Akihito Watabe, Health Specialist (Health Financing) at ADB. He continues, ‘AI can help to streamline claims management and detect patterns of fraud, waste, and abuse among healthcare providers through advanced data analytics and predictive modeling.’

 

 

Insights gained from the two Learning Exchanges held in 2025 are helping to ensure that the e-Learning course responds to the needs of health insurance decision makers and practitioners in partner countries.

Nepal leads the field

Of the countries participating in the initial Learning Exchange, Nepal was the first to embark on the development of AI for its health insurance sector. The country is already unique in using the openIMIS software to manage a nationwide social health insurance program. The global openIMIS Initiative is supporting the development of an AI module that will further enhance programme management by automating the processing of around 50,000 insurance claims on a daily basis and detecting irregular patterns in the data. This AI module is integrated into openIMIS, making it adaptable for use in other countries and programmes – a feature that attracted significant interest from other teams participating in the Learning Exchange.

In fact, the Nepal team is already providing support for the adaptation of the AI module for Cambodia’s National Social Security Fund, which provides health insurance for formal sector workers. Reflecting the strong spirit of collaboration fostered through the Learning Exchange, Purushottam Sapkota, Technical Adviser at GIZ Nepal, says, ‘We are here and ready to help others in their AI journey.’

 

 

From Learning Exchange to Learning Platform

Amref’s Institute of Capacity Development (ICD) is leading the transformation of the Learning Exchange materials into an accessible e-learning course format that will be freely available through Amref’s Jibu platform – a regional hub for digital health learning. It can be accessed via both mobile phone and laptop.

 

 

Reflecting on these materials, expert facilitator Tim Ohlenburg explains, ‘We’re developing new and cutting-edge content for the Learning Exchange that is different to that of traditional online courses – we aim to foster deeper engagement.’

The Institute’s Deputy Digital Learning Director, Willie Ngumi, who is spearheading the development of the new course, aims to generate a lively and interactive experience for learners. For example, he is experimenting with using AI-generated audio to replicate the experience of the Learning Exchange participants. ‘You should get the feeling you are in the room listening to a live lecture,’ he explains. Gamification – the use of quizzes, competitions and other games as educational tools – will help to embed learning and to make the process more enjoyable.

 

Structure of the e-learning course, AI for Health Insurance
 

 

 

A valuable opportunity for in-person exchange

The interactive workshop in Nairobi was the highlight of the first Learning Exchange for many participants. ‘This dynamic, in-person exchange allowed for deeper engagement, peer learning, and collaborative problem-solving,’ notes Amref Executive Director, Dr Rahul S Reddy Kadarpeta.

Over three days, participants from the six countries worked in small, mixed groups, and presented back to the plenary, ensuring that everyone had a voice. Describing the trust that had built up during the online sessions and which continued into the workshop, Rasheeda Buba, Principal Programme Analyst in the ICT Department at the Kaduna State Health Insurance Agency (NHIA) in Nigeria, says,

We were encouraged to discuss and to listen to the experiences of other countries. You could say the wrong thing and the facilitators appreciated this. This opportunity to make mistakes in a safe space is rare.

Dr Kadarpeta also highlighted the value of having both policymakers and technical experts in the same room, enabling grounded discussions that linked ambitious policy objectives with implementation realities.

Keeping the momentum going

The idea behind the Learning Exchanges is to create an online community of practitioners who feel comfortable working together – sharing mistakes as well as triumphs – while developing their capacity to work with AI. For many of the participating countries, the Learning Exchange represents the start of their journey towards integrating AI into health insurance management; Amref and GIZ therefore recognise the importance of keeping the community of learners active and connected. ‘Together they will go faster and develop further than by themselves’, says Simona Dobre.

Amref has established a LinkedIn group where learners and facilitators from both Learning Exchange cohorts can interact with each other, and new learners who access the e-Learning course will be given access to this group. If they have not already done so, participants are also encouraged to join the different committees of the openIMIS Initiative, including the Developers Committee and Implementers Committee, where they can join discussions, seek answers to technical questions, and access peer support. In addition, in early 2025 GIZ established an AI for Social Protection Hub that will offer guidance to governments interested in integrating AI into the management of social protection programs.

Following the Learning Exchanges, participants are well positioned to be emissaries for AI in their own countries and to build momentum for change among key stakeholders across sectors. Mr Senchheang Chun, ICT and Digitalization Advisor to the GIZ-managed Improving Social Protection and Health programme in Cambodia, says, ‘Now I understand better the journey of AI development; it’s not just about writing code or deploying systems, but a product development lifecycle that takes into consideration data governance, quality, and ethics, requiring broad knowledge of AI by many different stakeholders.’

 

Originally published by Healthy DEvelopments - Corinne Grainger, August 2025

For more information, please see the accompanying LinkedIn post.