Healthcare and life sciences have struggled to deliver quality outcomes for patients at scale. There have been three key binding constraints: a) a shortage and inelastic supply of good-quality healthcare professionals, b) a focus on curative rather than preventive and holistic care, and c) the complex and lengthy nature of drug discovery. This has meant poor healthcare access, rising costs, and unsolved diseases that have compromised the overall quality of life.
Artificial Intelligence (AI) is set to hit these binding constraints, with the potential to unlock outcomes that can make millions of people live longer and healthier. This is why healthcare is one area where the impact of AI can be most transformative, as it can elevate the overall human potential and wellbeing.
Recent developments of AI in healthcare demonstrate how it is addressing these challenges:
The recent rapid advancements in AI is a result of three forces coming together: advancements in compute with multifold decline in cost, copious availability of healthcare data and algorithmic developments.
Till the advent of the 21st century, the role of AI in healthcare was nascent and adoption was miniscule. The conceptual framework and foundational work for AI was laid in the 1940s. Two scientists, McCulloch and Pitts created a basic model of how the brain works in 1943 as the first idea of neural networks. The term AI in itself was advanced in the Dartmouth workshop in 1956. Through the 1960s the focus was on creating ‘expert systems’, that were computer programs designed to think like human experts that used IF-THEN rules (e.g., "IF fever AND rash, THEN consider infection"). Simultaneously through the 1980s, neural networks were developed that tried to mimic how the human brain learns by recognizing patterns from data rather than relying on predefined rules (IF-THEN). Early application of this in healthcare was reading ECG signals, analysing medical images from CT scans, MRI etc. However, computational cost and data restricted the scope of adoption of AI. It was in early 21st century that the key bottlenecks were unlocked:
1. Massive generation of healthcare data: The uptake of Electronic Health Records (EHR) in the USA picked up rapidly, from just 18% of physicians using EHRs in 2001 to 78% by 20134. The Health Information Technology for Economic and Clinical Health (HITECH) Act in 2009 in the USA provided financial incentives for meaningful use and adoption of EHRs. Suddenly we got a rich yet unstructured repository of clinical information from lab analysis, images, physician notes etc. The completion of the Human Genomics Project in 2003 (accounting for over 90% of the human genome) also became a fundamental block of personalised medicine and drug discovery later5.
2. Democratised access to computational power: The development of Graphics Processing Units (GPUs) and availability of GPU via cloud platforms democratised the access to high performance computing for training larger and more complex data sets. Further, the cost of compute massively went down; in a recent statement Jensen Huang, CEO of Nvidia stated, ‘in the course of the last 20 years, we've driven the marginal cost of computing down by one million times’6.
3. Algorithmic breakthroughs: The deep learning breakthrough happened post 2010s. In 2014, Ian Goodfellow introduced Generative Adversarial Networks (GANs), which allowed machines to generate realistic images. Google's 2017 paper ‘Attention Is All You Need’ introduced the Transformer architecture that changed how AI processes sequential data like text. The core innovation is a self attention mechanism that lets the model focus on specific parts of the sequence while making predictions. This was the backbone of later advancements in Large Language Models (LLMs) (e.g. Open AIs GPT series). Today, Generative AI spans multiple modalities such as images, speech, video, and coding.
Unlike other large industries, historically, Healthcare’s adoption of tech has been abysmally low due to multiple human, financial, and technological constraints. However, we believe that this is set to change now.
1. It’s no more garbage in and garbage out: There is no dearth of data; ~30% of the world’s data volume is being generated by healthcare7, albeit from multiple, disparate unstructured sources- EHRs, lab reports, pharmacies, insurance, wearables etc, sometimes being adopted in piecemeal ways by the healthcare providers. On top of it, there are multiple vendors often digitising data in non-standard formats. With data residing in silos, it is difficult to get a holistic picture of a patient's health journey that leads to suboptimal health outcomes. With advance Natural Language Processing (NLP) and Machine Learning (ML) tools, we can digitise, structure and analyse unstructured sources of data to make sense of the larger picture. Data privacy issues have also been addressed by federated learning where AI is trained on local servers without the need to share it at some centralised location.
2. AI’s adoption will be easier at scale: The healthcare industry’s adoption of technology and IT has been notoriously low. Some studies indicate that IT spending by healthcare, as a percentage of revenue, is ~5%, one of the lowest among large industries8. Healthcare professionals have been one of the biggest barriers to tech adoption, partly driven by the fact that traditional SAAS solutions required the last mile workers to change their workflows, do manual data entry time intensive tasks thereby adding to the list of their existing work streams. AI has the potential to automate a lot of grunt and manual work with limited need of training and behavioural change. For example a doctor may not need to take detailed patient notes or ask a nurse to digitize them; instead, a simple AI solution can record the interaction, generate detailed notes, and digitize the data in a structured format. This makes the lives of practitioners easy, and if implemented well can potentially unlock large scale AI adoption.
3. AI addresses the workforce constraints: Healthcare industry has always faced a shortage of high quality professionals including doctors, nurses and other clinical staff. The near term supply of healthcare professionals remains extremely inelastic and costly. The potential of AI is two fold: a) replace highly automated/structured workflows in the value chain thereby b) freeing up the time of professionals to focus on delivering solid clinical outcomes and potentially a greater capacity to serve more patients.
This is also one of the reasons why AI has the potential to not just disrupt the SaaS market, which is in the order of hundreds of billions of dollars (300+ SaaS unicorns in the last 20 years), but also the much larger services (human labour) market which is in the order of trillions9. Simply put, AI has the potential to replace or augment both software and human labour.
The USA is leading the way globally towards funding AI solutions in healthcare. In 2024, healthcare companies secured $23 billion in funding and 30% of this funding was directed toward startups leveraging AI10. High funding is a lagging indicator of favourable healthcare AI regulatory tailwinds. As of 2024, a total of 1,016 AI/ML-enabled medical devices have received FDA approval. Interestingly, only 335 of these were approved over the first 25 years (1995–2020), while in just the last four years, that number more than doubled with 681 new approvals11. Most approvals have come in Radiology (76%), followed by Cardiovascular applications (10%).
While in India the adoption of AI in healthcare is still nascent, we are seeing the large incumbents tinkering with AI. The healthcare IT incumbents are increasingly adopting AI and bundling it in their existing offerings. Innovaccer recently launched "Agents of Care," a suite of pre-trained AI agents designed to automate repetitive tasks for care teams12. Citius Tech has also launched Medictiv, an open healthcare AI model directory with a curated list of 250+ ready-to-use healthcare AI models13. IKS has made strategic investment in Abridge to develop accurate and trusted generative AI models for healthcare14.
Major Indian hospital chains like Apollo Hospitals and Max Healthcare, along with leading pharmaceutical companies such as Dr. Reddy's Laboratories are adopting AI to enhance their operations and improve patient outcomes. Apollo Hospitals has rolled out an AI-powered patient monitoring system that has led to an 80% reduction in Code Blue incidents and a 70% reduction in nurses' workload for recording vitals15. Dr. Reddy’s is using AI for accelerated drug discovery: AI is able to extract information from more than 2 billion compounds at a speed of 5,000 records per second, reducing a process that previously took months to less than five minutes. Its platform ensures high data quality, achieving >95% on AI data quality benchmarks, and exhibiting zero pipeline failures, saving an estimated 20% of monthly engineer effort previously spent on remediation16. They have also been able to achieve a 43% improvement in manufacturing cost, a 30% reduction in production lead time, a 41% reduction in energy consumption, and a significant decrease in quality deviations17.
We are seeing the entire healthcare journey of patients now being touched by AI from nonclinical administrative use cases (e.g. Revenue Cycle Management, scribing etc) to core clinical operations (e.g. diagnostics, patient monitoring and rehabilitation).
The relatively nascent Indian startup ecosystem is replete with examples where young, driven individuals and companies have disrupted incumbents. Consider how companies like Zerodha and Namma Yatri have challenged the industry giants to adopt newer business models. Today, even Uber and Ola are being pushed toward zero-commission or flat-fee models to stay competitive. While the notable Indian Health-tech companies like IKS Health, Indegene, and CitiusTech have grown and scaled globally, we strongly believe that the next wave of innovation will be driven by the agile, AI first startups.
New-age startups will attract top-tier talent, have a strong pulse on the market, and leverage this talent to accelerate the adoption of AI in healthcare. While we do have a broader investment criteria, we strongly look out for the following strengths in startups building in this space:
1. Deep understanding of healthcare: Healthcare systems are complex, marred by competing incentives of different stakeholders. Unless the team understands the major pain points of doctors, nurses, and other professionals and demonstrates how their solution is able to make the lives of these stakeholders significantly easier, they will struggle with adoption at scale. We want to know the nuanced understanding and the unique knowledge advantage that startups bring to the table.
2. Data edge: There are multiple sources of healthcare data, ranging from publicly available datasets to proprietary information data that startups may have access to through their operations. A few startups have also acquired data via strategic partnerships and investments with large healthcare providers and life sciences companies. Depending on the use case, the quantum and quality of data required may vary, so startups must have a clear, targeted strategy for data acquisition and usage aligned to their offering.
3. Distribution: Given issues around data privacy and stringent regulatory compliance, cracking early distribution in healthcare is hard. Founders who are already embedded in the system often use their first degree connections to get foot in the door, build credibility through outcomes and then scale. Startups like Hippocratic AI have taken a strategic approach by bringing large healthcare providers as investors from day one to co-build solutions that unlocked both access to proprietary data and large distribution for them.
4. Show outcomes, not features: Healthcare companies care about outcomes and ROI: a) demonstrate tangibly how your solution is increasing revenue or reducing costs, and b) making the lives of health professionals easier. More features will not necessarily add any incremental value unless your solution is able to demonstrate the above outcomes.
Interestingly, we are seeing healthcare companies adopting outcome based pricing models globally. For example, Sword Health (founding year 2015, USD 3 bn valuation18), provider of digital therapeutics for physical rehabilitation, has introduced a new outcome-based pricing19, charging clients only when a participating employee's health shows improvement with better engagement.
At Enzia, we are excited about the following opportunities:
A. Horizontal agentic AI platforms: We are excited about agentic AI platforms capable of automating multiple healthcare workflows, from patient enrollment to post-discharge care. However, a key consideration is to handhold the client in this journey. Specifically in healthcare with stringent regulatory requirements, behavioural challenges and data interoperability issues, a one-size-fits all solution will be difficult to implement and scale. Therefore, supporting healthcare organisations through this journey with seamless integration in the existing systems would be the key to success.
B. AI-first health services company: An opportunity exists for an AI led health services company to not just remove barriers to tech adoption through AI but also democratise access by going beyond catering to the fortune 500 companies and expanding their offerings to newly funded startups that find the services of large IT incumbents cost-prohibitive.
Startups like Reveal HealthTech (founding year 2023, raised ~USD 8 mn) are prime examples. With a stellar team based in India and USA the company has developed novel AI solutions across clinical and non-clinical use cases and is already serving both PE funded startups and fortune 50 companies in the USA.
We see significant potential of AI serving point solutions targeting large profit pools. Revenue Cycle Management (RCM), which optimizes the financial flow from patient to provider is a case in point. In USA, of the USD 222 bn TAM for healthcare provider-centric technology solutions, roughly 46% ie. USD 102 bn belongs to RCM alone, representing a critical workstream where providers currently allocate 3.7% of their revenue20. India has been the backbone for several large RCM companies with revenues exceeding ~USD 360 mn. Today, the emergence of AI-first RCM players presents a significant opportunity to leapfrog traditional models and deliver greater efficiency, accuracy, and scale.
Startups such as Rapid Claims (founding year 2023, raised USD ~11 mn) have grown 6x within the year of their launch with stellar outcomes delivered (>98% Clean claim rate, 40% reduction in denials with a 70% reduction in cost to collect).
Given the massive shortage of healthcare professionals, the most disruptive and needle moving use of AI would be as a force multiplier for the existing healthcare professionals. This is where a robust CDSS system will come into play that leverages the data available from all the sources and helps generate timely and informed advice/diagnosis at the point of care.
India offers a distinct competitive advantage for AI development in healthcare, driven by its large and diverse patient populations, cost-effective data labeling and processes, and access to clinical expertise. Companies like Qure.ai have effectively leveraged this by training their initial models on extensive local datasets and fine-tuning them for global deployment across varied demographics. Looking ahead, we anticipate therapeutic focused companies (e.g. neurodivergent children, dementia, gut health) from India to leverage rich data and AI to build robust personalised clinical protocols; and AI-led diagnosis opening up possibilities in other therapies. We expect these companies to expand their reach to international markets with larger revenue pools to support clinicians in better decision making, further strengthening the global impact of India’s healthcare AI ecosystem.
We are already seeing AI based diagnosis and automation companies at the forefront of innovation. Cancer diagnostics is a USD 136 bn market growing at 8% annually. Histology remains a major chokepoint in terms of manpower costs ($25-35/hour 5 years back has increased to $35-$50 in 2024) and availability (6% vacancy rate in Histology in 2017 has grown to 13.2 % in 2022)21. For over a century, histological tissue analysis has been the gold standard for confirming cancer diagnoses and informing treatment plans. However, despite technological advancements in other areas, histology has seen virtually no automation gains in the past decade. As a result, approximately USD 10 bn is currently spent each year on manpower for histology procedures worldwide. In the histology workflow, processing and analysing tissue culture remains one of the bottlenecks given its manual nature and lack of skilled professionals in the USA. One of our portfolio companies, Morphle Labs is automating the main chokepoint in histology diagnosis mainly by automating microtomy (ultra-thin slices are cut from tissue blocks and transferred to glass slides and scanned using their other product, tissue scanner). Their hourly throughput beats fastest human operators by a factor of 2 today and is expected to improve further. Digitizing these images will enable the creation of a marketplace for AI algorithms, many of which have seen limited adoption due to lack of access to structured imaging data.
The USA healthcare market presents a massive opportunity for solutions built from India with a staggering USD 4.9 trillion in spending in 2023 (17.6% of USA’s GDP, $14,570 per person; 125% of Indian GDP) and projected to reach USD 7.2 trillion by 2031 (~$20,425 per person)22. On the back of our large successful IT services companies, India has developed proven playbooks to succeed further in this space.
One of our portfolio companies, Circle Health, offers proactive care through a dedicated Care Manager who serves as a personal health guide, providing 24/7 support for both planned and emergency hospital visits, as well as everyday health needs. Having established a successful playbook to deliver unheard of health outcomes in India- employee engagement rate of 70–80% (compared to the industry average of around 30%) and a claims ratio of 76% (versus the industry average of 100%)23 they are now expanding to the USA. Circle Health is also leveraging AI as a force multiplier to help care managers automate backend tasks and deliver care more efficiently to a larger number of people. They recently signed their first contract with a Skilled Nursing Facility (SNF) to offer improved, tech-enabled Remote Patient Monitoring (RPM).
AI can disrupt healthcare. The key will be to marry it with solutions that ease adoption, remove the fear of unfamiliarity, and reduce the cognitive load on health professionals to adopt AI at scale.
-----------
Endnotes
1. Kaviani, P., Digumarthy, S. R., Bizzo, B. C., Reddy, B., Tadepalli, M., Putha, P., Jagirdar, A., Ebrahimian, S., Kalra, M. K., & Dreyer, K. J. (2022). Performance of a chest radiography AI algorithm for detection of missed or mislabeled findings: a multicenter study. Diagnostics, 12(9), 2086.
2. Meltzer, E. (2024, November 14). Council Post: AI in Healthcare: A new era of Personalized Patient care. Forbes.
3. Viswa, C. A., Bleys, J., Leydon, E., Shah, B., & Zurkiya, D. (2024, January 9). Generative AI in the pharmaceutical industry: Moving from hype to reality. McKinsey & Company.
4. Use and characteristics of electronic health record systems among office-based physician practices: United States, 2001-2013. (2014, January 1). PubMed.
5. Human Genome Project Fact sheet. (n.d.). Genome.gov.
6. Lopez, J. (2025, January 9). Nvidia CEO Jensen Huang claims massive cost reduction. DevX.com.
7. RBC Capital Markets | Navigating the changing Face of healthcare episode. (n.d.).
8. Weins, K. (2025, February 4). IT spending by industry. Flexera Blog.
9. Yoo, J. (2024, August 15). Why Will Healthcare be the Industry that Benefits the Most from AI? Andreessen Horowitz. https://a16z.com/why-will-healthcare-be-the-industry-that-benefits-the-most-from-ai/
10. SVB's 2025 Healthcare Investment and Exits report
11. As on Sep’24, Enzia Analysis, Center for Devices and Radiological Health. (2025, March 25). Artificial Intelligence and Machine Learning (AI/ML)-Enabled medical devices. U.S. Food And Drug Administration.
12. Innovaccer. (n.d.). Innovaccer launches ‘Agents of CareTM’ to transform healthcare operations.
13. CitiusTech. (2021, August 3). CitiusTech launches Medictiv, a first-of-its-kind open healthcare AI model directory to drive healthcare industry collaboration. Business Wire.
14. IKS Health, & Abridge. (2023, May 18). IKS Health makes strategic investment in Abridge to launch a partnership that enables the creation of accurate and trusted generative AI models for healthcare. Business Wire.
15. AI-powered patient monitoring system at Apollo Hospitals cut Code Blues by 80%. (2025, February 6). Hospital Management Asia.
16. Informatica. (n.d.). Dr Reddy’s revamped cloud AI data strategy.
17. Dr. Reddy's. (2022, October 11). World Economic Forum Recognises Dr. Reddy's Hyderabad Factory as Part of its Global Lighthouse Network. Business Wire.
18. Tracxn
19. SWORD Health. (2024, September 10). Sword Health unveils outcome pricing: Takes Industry-Leading Approach to Pay for Performance. GlobeNewswire News Room.
20. Inventurus Knowledge Solutions Limited. (2024, December 5). Red herring prospectus.
21. Garcia, E., Kundu, I., Kelly, M., & Soles, R. (2023). The American Society for Clinical Pathology 2022 Vacancy Survey of medical laboratories in the United States.
22. NHE Fact Sheet | CMS. (n.d.).
23. Average Incurred Claims Ratio, Group Health Business at ~101% between FY 21-FY24
.png)
.png)
.png)
.png)
.png)
.png)