Artificial intelligence (AI) is poised to transform healthcare delivery, with diagnostic applications, particularly in radiology, representing one of the most mature and impactful domains. In South Asia, the potential for AI to bridge gaps in healthcare access and quality is substantial. India’s AI healthcare market, for instance, is projected to expand from US$0.95 billion in 2023 to US$6.5 billion by 2030.1 This growth is driven by factors including an aging population, rising incomes, and government-led demand through initiatives like Ayushman Bharat.2 However, a significant disconnect exists between this potential and the reality on the ground, where approximately 68% of clinical AI deployments remain in pilot or proof-of-concept stages.1 The transition from pilot to scalable, sustainable implementation is hindered by a complex web of interconnected challenges.
The core problem this paper addresses is the lack of a holistic framework for understanding what enables low-cost diagnostic AI platforms to succeed in the unique socio-economic and regulatory context of South Asia. While technological advancements are rapid, their adoption is contingent upon viable business models, acceptance by key stakeholders (patients, clinicians, hospitals, and payers), and navigation of an evolving regulatory landscape. This study aims to analyze these interdependent factors to identify pathways for sustainable growth and scalability. By synthesizing findings on revenue models, adoption barriers, and regulatory frameworks, this paper provides a comprehensive case study of the challenges and opportunities for diagnostic AI in the region.
The successful deployment of diagnostic AI platforms is contingent on demonstrating clear value, overcoming systemic barriers, and designing for the complexities of a multi-sided market. The literature points to significant potential, but also to formidable obstacles that are particularly acute in the South Asian context.
AI applications in healthcare offer a range of value propositions, including risk-reduced patient care, process acceleration, resource optimization, and knowledge discovery.3 In diagnostics, AI has proven capable of reducing the workload of healthcare workers by over 30% while maintaining accuracy, a critical benefit in regions with shortages of skilled professionals.4 Economic models further project significant long-term cost savings from AI integration, particularly in treatment pathways informed by more efficient diagnostics.5 These benefits form the core value proposition that platforms must deliver to stakeholders.
Digital health platforms operate within complex ecosystems and function as multi-sided markets, requiring the creation of mutual value for all participants.6 A successful platform design must begin with a focus on these multi-sided value propositions, which can be structured hierarchically: first-order value serves patients, second-order value serves care providers (e.g., doctors, hospitals), and subsequent orders cater to developers and the platform owner.6 This perspective underscores that a singular focus on patient or clinical benefit is insufficient; a sustainable business model must also address the needs and incentives of payers, administrators, and technology partners.
Despite the clear potential, AI adoption in South Asian low- and middle-income countries (LMICs) is impeded by deep-seated structural challenges. A primary barrier is the lack of consolidated, high-quality data. Health data systems are often fragmented, with many hospitals still relying on paper records and having little incentive to digitize or share information.4,7 This data fragmentation is a fundamental obstacle to training and deploying effective AI models. Additional barriers include inadequate digital infrastructure, limited funding, a shortage of digitally skilled healthcare professionals, and a lack of established governance structures for AI.8 In India, specific challenges include poor workflow integration, where the data entry burden on health officers can undermine adoption, and the need for solutions designed for environmental constraints, such as offline functionality in lightweight apps for use in rural areas.9
This study employs a qualitative synthesis of existing research, drawing from peer-reviewed academic articles, government and regulatory publications, and specialized industry reports. The analytical approach is framed by a multi-sided market perspective, which is essential for understanding the dynamics of digital platform ecosystems. The research synthesizes evidence across three interconnected pillars that collectively determine the viability and scalability of diagnostic AI platforms in South Asia: 1) business and revenue models, 2) stakeholder adoption strategies and barriers, and 3) the regulatory and reimbursement environment. The analysis focuses primarily on India as a comprehensive case study due to the availability of data and its significant market size, while also drawing comparative insights from other Asian countries to highlight diverse policy approaches. By integrating these three pillars, the methodology aims to construct a holistic view of the factors driving or impeding the sustainable growth of low-cost diagnostic AI.
The analysis reveals that navigating the South Asian healthcare market requires innovative strategies that extend beyond technology to encompass business models, regulatory compliance, and reimbursement pathways. Success is contingent on aligning with existing systems while simultaneously addressing their inherent limitations.
Given the complexities of data privacy regulations and market fragmentation, digital health startups in Asia increasingly favor a Business-to-Business (B2B) model. This approach allows them to partner with established stakeholders like hospitals, which are already compliant with regulations, thereby reducing their own compliance burden.10 A particularly effective model in India is the Public-Private Partnership (PPP), where private enterprises provide capital and clinical services while the government acts as the payor. Manipal HealthMap Diagnostics exemplifies a successful PPP in radiology and pathology, demonstrating a viable path to delivering services at scale within the public healthcare system.2 Furthermore, large-scale telemedicine platforms such as Practo and Apollo 24|7 are integrating AI-powered tools, creating ecosystems that facilitate remote consultations and expand healthcare access, particularly for rural communities.4
India’s regulatory framework for AI-based medical software is evolving. The Central Drugs Standard Control Organisation (CDSCO) is the primary regulatory body.11 In September 2021, the CDSCO released guidelines for SaMD, harmonizing its risk-based classification (Class A, B, C, D) with the International Medical Device Regulators Forum (IMDRF) framework.12 Notably, the CDSCO has not yet identified any SaMDs in the highest-risk Class D category and requires SaMDs to follow the same registration process as physical devices in their corresponding risk class.12 However, the existing Medical Devices Rules (2017) do not yet contain specific provisions for SaMD or sophisticated channels for managing iterative AI/ML product updates.13 To address this, industry bodies have proposed frameworks like a “Predetermined Change Control Plan” to streamline the approval of algorithm modifications without requiring a full re-review for every update.14
A major hurdle for diagnostic AI adoption in India is the lack of explicit reimbursement standards for digital health technologies.15 To address this, the National Health Authority (NHA) established a Health Financing and Technology Assessment (HeFTA) unit to use HTA evidence to inform which technologies are included in the public insurance scheme, Ayushman Bharat PM-JAY.15,16 However, the use of economic evidence remains nascent. The 2022 revision of the Health Benefit Package was the first to incorporate HTA evidence, but its application was limited to oncology.16 Significant barriers persist, including a lack of timely HTA evidence and inadequate capacity among NHA staff to conduct cost-effectiveness analyses.16 A 2022 review found that no digital health tools, including AI-powered ones, were reimbursed in India, highlighting a critical gap between innovation and financial viability.15
Successful adoption strategies are tailored to local realities. One effective approach is integrating AI models into large-scale, existing government technology platforms. For example, an AI model to predict patient ‘loss to follow-up’ was successfully integrated into India’s national Nikshay platform for Tuberculosis management, leveraging its vast dataset and existing user base of health workers.9 Designing for specific user and environmental constraints is also crucial. An AI system that initially failed due to the high data entry effort required from health officers was successfully redesigned with a voice-based interface.9 Finally, the government is using financial levers to spur adoption. In 2023, the NHA launched a Digital Health Incentive Scheme, offering significant financial incentives to hospitals, labs, and tech companies to integrate with the Ayushman Bharat Digital Mission (ABDM) ecosystem.10
The findings underscore that for diagnostic AI platforms to achieve sustainable scale in South Asia, technological innovation alone is insufficient. Success hinges on a triad of strategic alignment: a viable business model, a clear regulatory pathway, and a reliable reimbursement mechanism. The challenges identified are not unique to AI but are amplified by its data-dependent nature and its role as a novel intervention category.
The core challenge is the misalignment between the value propositions of AI and the practical realities of the healthcare ecosystem. Data fragmentation prevents the development of robust models,4,7 while workflow integration issues hinder adoption at the provider level.9 Even when these are overcome, the absence of a clear reimbursement pathway makes it difficult for providers to justify the investment.15 Therefore, a sustainable business model must actively address these friction points. B2B and PPP models are gaining traction precisely because they embed the AI solution within an existing clinical and financial structure, whether a private hospital chain or a public health program, thereby mitigating some of the adoption and reimbursement risks for the developer.2,10
Comparing India’s developing framework to more mature regulatory systems reveals potential pathways forward. South Korea, for example, has established specific value assessment guidelines for AI in radiology and pathology.15 It also provides a grace period of up to three years for new AI devices to build clinical evidence for reimbursement, during which providers can charge patients out-of-pocket.17 Furthermore, South Korea’s regulator has a progressive policy that exempts certain performance-enhancing algorithm updates from re-approval if the developer has an established change management policy, fostering rapid, iterative innovation.18 Similarly, Malaysia has established “Health Technology Hubs” in five hospitals to create regulated sandboxes for testing and commercializing digital health innovations.10 These international examples provide models for creating more agile and supportive ecosystems that could be adapted for the South Asian context.
The journey of low-cost diagnostic AI platforms in South Asia from pilot to widespread, sustainable implementation is complex and multifaceted. This paper has demonstrated that success is not merely a function of technological sophistication but is critically dependent on the strategic integration of business models, stakeholder adoption, and regulatory frameworks. The primary barriers identified—data fragmentation, workflow integration challenges, and the absence of clear reimbursement pathways—require systemic solutions. Emerging business models like B2B and PPPs offer promising avenues for navigating these obstacles by aligning with existing institutional structures. However, for the market to truly mature, a more predictable and supportive policy environment is essential. Future research should focus on longitudinal case studies of platforms that successfully scale, quantifying the impact of specific business models and policy interventions. Furthermore, developing and institutionalizing robust HTA capacity within national health authorities will be paramount to ensuring that clinically effective and cost-effective AI innovations are integrated into public and private reimbursement schemes, ultimately unlocking their full potential to transform healthcare in the region.
