‘Picking the right AI technology is only one-third of the battle’

In the healthcare industry, the patient is at the centre. So, the most important aspect is the right and timely diagnosis of their issues and the right intervention at the right time. What prevents this from happening at all times? One of the key reasons is the lack of information available with different stakeholders; be it the patients, be it the caregivers, be it the prescribers…At ZS, we believe that AI – and especially generative AI – can make a very strong impact in bringing the right information available at the right time.

That is what is happening through chatbots, through intelligent web applications, through intelligent software medical devices. Coming closer to our home…India…I think given the constraints on the resources front, the number of providers and the number of caretakers – disease prevention and early diagnosis are extremely critical. If we look at any of the non-invasive approaches like retina scans or using saliva, and then leverage AI to understand if we can diagnose any issue early on in the lifecycle…that will have a big impact on patients, especially those who do not have access to the labs. The second aspect within the patient diagnosis world is the right diagnosis and that’s where AI is making a big impact by mining through a lot of data. The other area we at ZS support many clients is R&D. For the healthcare industry, R&D is a negative ROI proposition to begin with. For instance, if 100 molecules are put in the pipeline, it’s only four molecules that get to see the light at the end of the day. So, the throughput is just 4 per cent. That’s where AI is making a big impact, by improving our ability to target the right molecules. So instead of 4 per cent, we are actually able to see 12-14 per cent throughput. Instead of taking 6 to 8 years in completing the clinical trials, as we saw in the Covid vaccination experience as well, AI is playing a big role in identifying the right sites, identifying the right set of patients and improving the efficiency of clinical trials.

My strong belief is that picking the right AI technology is really all about looking at it first from a business-first mindset. It should really start from developing a very deep understanding of the business model today and the future of the business model. It is critical to recognise that AI technology is evolving at a very brisk pace. So, you pick a suite today, and in two months, the suite could look very different and very new. So how we go about helping our clients is first focusing on the foundation skills, which is having the right data and having the right business process. Then enabling the two by picking the right AI technology. But picking the AI technology is only a third of the battle because you don’t want to get stuck with any particular technology for the long term. Hence, depending on the use case, it could be that you go with a particular AI technology with a very strong foundation and then keep adding on different layers to enable your end-to-end business outcome. 

AI, including generative AI, is a pivotal piece of this larger puzzle. It is not an isolated technology but is indeed part of a continuum that is shaping the future. And the way I see it, the best AI is invisible. And what I mean by that is that it’s not really AI, but what matters is the experience that it creates, the decisions it enables us to make and the tasks that it helps us accomplish.

Sometimes the decisions are very small and hence just one technology may be enough. However, many decisions, especially in healthcare, are very complex. And so, we are increasingly seeing that it is the continuum that truly plays an important role, and you need to bring multiple aspects of AI together. These include classical AI, generative AI, the digital aspects of it, machine learning, deep learning, etc. For example, classical AI can help you identify the right patient who might be in need. Generative AI can then help you with the draft on how you would engage that patient to get him or her on board with your support programme. Maybe tomorrow, Metaverse will play an important role in engaging the patient and actually helping them out remotely. And so, we see it more as a continuum.

Adoption is a challenge, but some recent examples tell us that it can be overcome. I’m sure you are using UPI yourself. So that’s a technology that was introduced not too long ago and now in two or three years, we are using UPI for majority of our financial transactions. Likewise, DigiLocker has really taken a good shape and is being adopted very widely. This basically tells us that if we are able to reduce the barriers for people to adopt, our country is absolutely ready to be at the digital forefront. And so, I feel that the roadmap really includes having the right set of partnerships, having the right set of funding and then looking through the customer lens. If I were to break it down, the first part, which is the right funding and the right mindset, I must congratulate our Central Government which is being really relentless in terms of bridging this digital divide. Look at the National Digital Health Mission (NDHM), which is enabling many of the start-ups to leverage the data that is being made available for them to make the best use of AI technologies and create new functions.

We’ve also seen the government increasing the health expenditure by over 2.5 per cent in the budget, which is great. I personally believe that the biggest hurdle for any adoption is trust. How do we earn trust? It is all driven by privacy of the data, having the right proof points, and making it a lot more accessible so that people can try and share their stories. The more stories people hear, the more trust they gain. So, largely speaking, my focus is really to see how organisations can help gain more trust, and once the trust is gained, I see no issue in the adoption of these technologies.