Blog

Artificial Intelligence (AI) has proved to be a gamechanger in a range of fields – from creating personalised shopping experiences for retail customers, bringing automation to financial services, and enhancing diagnostics and patient outcomes in healthcare.

According to the World Economic Forum, the economic potential of generative artificial intelligence (GenAI) is immense. Goldman Sachs states GenAI could boost annual productivity by 1.5%, driving $7 trillion in added economic value over the next decade while McKinsey predicts that GenAI could add up to $7.9 trillion annually to the global economy.

Powering these advancements is an onslaught of data – massive in volume, complex in nature, and generated at a high velocity. Together, AI and data have transformed the world we live in, with strategies centred on data management and storage providing organisations with a blueprint to manage, leverage, and harness data assets. Armed with these strategies, organisations at the forefront of change can ensure that the right data is collected and stored securely, and made accessible and available for analysis and decision making.

However, the generational shift of Generative AI (GenAI) brings a question: how can we build future-proof strategies for data and Generative AI? How will traditional strategies keep up with this paradigm shift?

Industry leaders from various sectors came together to discuss the potential and pitfalls of data and GenAI in a closed-door roundtable organised by Couchbase and Google, and hosted by YourStory Media.

Participants spoke at length about their journeys with GenAI, discussed industry-specific insights on the technology, identified risks such as hallucinations, spotlighted potential use cases, and weighed in on the business Return on Investment (ROI) of GenAI.

The roundtable featured an impressive gathering of thought leaders and experts from a range of industries: Anant Pal, Head – Product, Propelld; Ravindra Yadav, Senior Director – Data Science, Meesho; Puneet Tripathi, Head – Data Science, Wakefit; Hemant Misra, SVP & Head of Data Science, Simpl; Suraj Yadav, Head of Data Engineering, Zee Entertainment Enterprises Limited; Mallesh Bommanahal, Chief Data Scientist, ACKO; Dhiraj Prakash Naik, Senior Director – Analytics, Razorpay; Amit Chandel, CTO, Olyv India; Ankit Chaudhary, Head of Data Practices, Shadowfax; Praveen Singh, Senior Director- Data Engineering, PayU; and Vishal Ramrakhyani, VP and Head of Engineering, ZoomCar.

The panel began with short keynote addresses by Krishna Thirtha, Regional Business Head, Couchbase, and Debasis Bhattacharya, Head of Customer Engineering, Google Cloud, and was moderated by Shivani Muthanna, Director – Strategic Partnerships & Content, YourStory Media.

Key takeaways

The discussion explored the various use cases for AI and GenAI, including creating AI assistance for loan underwriting, leveraging generative AI with models like Google Gemini or OpenAI for unstructured data, and providing text chatbots and voice bots for enhanced customer support. Panellists concurred that customer support was the predominant and most successful use case for GenAI.

Assessing the business ROI for GenAI was an important point of discussion for panellists. While companies have reported varied levels of success with GenAI tools, the ROI on these investments remain murky. Some issues that have cropped up with the continued use of GenAI include affordability, hallucinations, and model drift (the decay of machine learning models over time due to changes in data). Companies continue to be cautiously optimistic about the business ROI, opting for a wait-and-watch approach at present.

A key insight uncovered during the discussion was the size of GenAI models. Initially, there was a fever pitch around building large models that would solve multiple problems simultaneously. However, with time and experience, model providers have started offering models of various sizes. This gives companies an opportunity to leverage Retrieval Augmented Generation (RAG) techniques, and fine-tune their models to address targeted problems.

Organisations can often lose themselves in the vast amounts of data generated everyday. Untapped, this data can remain in silos, causing paralysis, indecision, and decreased productivity for employees. This issue – known as a data sprawl – was discussed at the roundtable, with experts highlighting the large scope of the problem. Panellists advised companies to look for data platforms actively working to aid employees in mining this data in an organised and efficient manner.

Speakers also examined GenAI through the lens of security, privacy and consent. Sensitive and intimate information, such as Personal Identifiable Information (PII) and Protected Health Information (PHI), must be secured to prevent data leakage. Participants urged the use of data anonymisation, as well as practising access control – making certain data accessible via approvals from top management.

GenAI has clearly reported success in arenas like customer service. However, businesses continue to tread lightly around adopting this technology. The discussion highlighted how businesses are opting for a slow-and-steady approach to GenAI, as they continue to discover new use cases and uncover its true ROI.