Telkom University and AirNav Launch Offline AI School Tool

The flagship product, Edunusa, is designed for elementary students and differs from mainstream chatbot-style AI by using a Socratic method that nudges children toward answers instead of handing over ready-made solutions. That approach is intended to address one of the fastest-growing concerns in education: students using generative AI to complete assignments without learning the underlying concepts.

Prof. Agus Pratondo, director of Telkom University’s Center of Excellence Artificial Intelligence, Machine Learning and Optimization, said community education programs in remote areas are often prohibitively expensive because lecturers must travel in, train teachers for a short period and leave, after which updating material becomes difficult. With Edunusa, instructional content is stored on local servers and updated centrally when connectivity is available, then pushed back to remote sites.

For investors and policymakers, the significance is broader than one product launch. Indonesia’s digital inclusion agenda has repeatedly run into uneven broadband coverage, high last-mile costs and the challenge of training teachers at scale. A server-based AI model offers a way to bypass part of that bottleneck and could become a template for other public services in remote regions, including health education and vocational training.

The educational angle also carries a governance benefit. Because the system does not require open internet access, it can reduce exposure to harmful online content and tighten data control, an increasingly important issue as schools experiment with AI tools that often rely on third-party cloud infrastructure.

Telkom University also introduced Mediciana, an AI model aimed at medical students, and SafeMe, a framework to assess AI maturity and ethics across institutions, suggesting the university is positioning itself as a domestic builder of what Indonesian officials increasingly describe as “sovereign” AI — tools developed locally, tailored to national needs and less dependent on foreign platforms.

AirNav Indonesia’s role underscores another important economic point: the companies involved are framing AI as infrastructure for efficiency, not as a labor substitute. Zainal Arifin Harahap, AirNav’s technical director, said technology should be adopted only when it clearly supports safety and work processes, a message that could shape how state-linked firms justify AI spending in Indonesia’s public sector.

The government is signaling support. Muhammad Fauzan Adziman, the director general for research and development at the Higher Education, Science and Technology Ministry, said the projects align with a national priority to lift participation and quality in 3T education, adding that the ministry could help finance further development through universities.

For education technology investors, the bull case is that offline AI solves a real structural problem in emerging markets: how to scale quality instruction where bandwidth, budgets and teacher supply are all constrained. The bear case is that adoption in the field will depend on maintenance, local training and sustained funding, and even an offline AI system still needs governance, content updates and hardware deployment to work at scale.

The next test is whether the pilot can move beyond showcase status into a repeatable model across remote schools. If it does, Indonesia could be laying the groundwork for a cheaper, more resilient version of AI-assisted public education — one built for the parts of the country that digital infrastructure has reached last.