Industry experts weigh in on leveraging AI to make better sense of data
In a panel discussion at TechSparks 2023, industry experts shared their experiences in handling data and the role of AI in unlocking its true potential.
Data is often compared to precious resources like oil. However, the truth is, data is more like water - it's everywhere, in vast quantities. In today's fast-paced, data-rich world, the ability to extract meaningful insights from vast datasets is a critical driver of success for businesses and organisations.
In a panel discussion during the recently concluded TechSparks 2023 on the topic ‘Making better sense of data using AI’, industry experts converged to share their experiences in handling data and the role of AI in unlocking its true potential.
The panel featured Ankit Maheshwari, CTO & Founding Member,
; Kaushal Shubhank, VP, Engineering, ; Hilal Ahmad Lone, CISO, ; and Animesh Bansriyar, Head of Solutions Architecture, Emerging Markets, APJ, Elastic. The session was moderated by Madanmohan Rao, Research Director, YourStory Media.Taming the data ocean
Kickstarting the discussion, Bansriyar highlighted the ubiquitous nature of data challenges. Across industries, from ecommerce to streaming platforms, the fine-tuning of AI models to cater to domain-specific contexts remains a pervasive hurdle.
“At each stage of the process, there are distinct issues that require attention. This encompasses defining the specific business problem at hand, ensuring data is appropriately prepared, and training AI models suited to the task. Verification of results is also paramount. Finally, the deployment of these solutions into production and making real-time business decisions are pivotal,” he said.
Maheshwari shed light on the monumental impact of AI on the US healthcare market. By harnessing vast troves of patient data, totalling 20 to 30 million records, Innovaccer predicts patient outcomes and empowers providers to deliver more cost-effective care.
"One impactful instance that comes to mind involves using our data to anticipate the risk of patient re-admissions in the US. This is a critical aspect, as it aims to prevent patients from being readmitted for the same issue repeatedly, which incurs significant costs on the healthcare system. By analysing a patient's historical data and considering various factors, we can predict the occurrence of multiple chronic conditions. This proactive approach not only lowers healthcare expenses but also enables providers to offer improved care to patients,” he shared.
Operating in the high-stakes fintech industry, Lone grapples with the intricacies of handling massive transactional data. The fintech major deals with substantial volumes of transactional data, encompassing sensitive information like Personally Identifiable Information (PII), card data, and financial records. This data is inherently susceptible to potential misuse, making it imperative to combat fraud and abuse.
“Even before the surge in popularity of generative AI, we employed AI and Machine Learning to conduct risk analysis, quantifying potential risks associated with our data. This has been a longstanding practice that predates the advent of generative AI. We've dealt with colossal volumes of data and AI has emerged as a crucial tool for us," Lone said.
Shubhank discussed the challenges of managing data from their lockscreen product, which is on 200 million devices in India. He highlighted the complexity of providing personalised content and interactions on a device's lock screen.
“On average, people unlock their phones about 150 times a day, indicating the extensive data volume they have access to. Sometimes, users unlock their phones but may not necessarily engage with Glance. They might consume various content like games and entertainment. Given this interaction data, the aim is to discern users' preferences and serve them accordingly, taking into account the unique real estate of the lock screen,” Shubhank explained.
Unlocking the true potential of AI with personalisation
In today's digital landscape, personalisation is integral to any search experience. It involves understanding user behaviour, including past purchases and demographic information. This personalisation aspect is now a fundamental feature in various ecommerce and SaaS platforms, where user profiles play a key role.
Bansriyar acknowledged that while there have been efforts to address personalisation challenges, especially in industries like ecommerce and streaming platforms like Netflix, a comprehensive solution is yet to be fully realised.
Shubhank revealed that personalisation is central to Glance's offerings. The platform provides a diverse range of experiences, including live entertainment, gaming, news, and shopping. The challenge lies in understanding how users engage with these experiences and what preferences they might have across different domains. It's not merely about one interaction leading to a similar one, but rather comprehending the broader spectrum of user preferences. This complexity underscores the significance of personalisation.
Shubhank acknowledged that there's still much work ahead to ensure users are engaged in a meaningful way, andthat Glance is still at the initial stages of this process.
The panellists demonstrated how AI-driven data sensemaking isn't just a technological feat but a strategic imperative. From healthcare to finance, entertainment to ecommerce, the transformational power of AI is reshaping industries and paving the way for a future where data isn't just abundant but invaluable.