“India has no dearth of data,” said IBM’s country leader – data, AI and automation (India and SE), Siddhesh Naik, in an exclusive interview with Analytics India Magazine. 

Citing the government cancer hospital in Mumbai, Naik said that the country probably has data much bigger than some of the medical record giants in the world. The diverse DNA data available here can fuel oncology research in all probability. However, the biggest challenge here is how to channelise all this data. 

In this regard, the government is already taking baby steps. For instance, Minister of State Rajeev Chandrasekhar recently said that the government had drafted a policy to share anonymised data sets collected under the National Data Governance Framework with Indian startups and researchers. The policy will standardise all undefended data, data management, and data processing silos—how to store, manage access, control and security—such that they will constitute an administrative component of the policy and its first pillar. 

“The government will get out all the data to industries as an API in a secure manner (while ensuring all the privacy requirements are met) so that large enterprises and startups can build their offerings on top of the API,” shared Naik, saying that IBM would focus on building a strong technology foundation. For example, India is currently going through a wave where there is too much excitement over developing data science models. However, he believes this can only be sustained if there is work done on data engineering. 

Further, the move to multi-cloud and hybrid environments has pushed enterprises to work towards bringing data at arm’s length to all departments within the organisations. This has led to data democratisation—making data available as low as possible within an organisational hierarchy chain—being an essential asset to enterprises. 

The future is AI governance 

Giving the example of a painting giant, Naik said that it is working with them to democratise data, where they are helping them in having the right governance—in terms of what can be exposed and what cannot—as well as sufficient data to be available for analysis. So, for example, the marketing team can utilise it to bring campaigns accordingly. Since, when it comes to paints, what works in one region doesn’t necessarily work in another. Hence, data democratisation is about hyper-personalisation/hyper-localisation, personalising data across all levels of the organisation. This way, IBM is working on bringing down the centre of gravity and making data consumable at a much deeper level.  

But applying it continues beyond the AI part. One has to sort the entire execution workflow process that enables it. “The beauty about it is that we are moving at an unprecedented pace where there is no resistance to getting there; everybody’s clear that we will get there and have started taking baby steps,” added Naik. 

IBM is addressing the challenge of data democracy with technologies like data fabric. As defined by NetApp, Data Fabric is, “an architecture and set of data services that provide consistent capabilities across a choice of endpoints spanning hybrid multi-cloud environments. It is a powerful architecture that standardises data management practices and practicalities across cloud, on-premises, and edge devices”. Thus, it is a virtual fabric that connects all data endpoints. Here, you use technologies like automation, federated governance, integration, security, and all of that, to be able to harness the data without physically moving it to have the ability to create an analytics backbone or dashboard that cuts across all silos. 

Another example is how with 5G coming in, the partnership of IBM with the telecom giant Airtel enables consuming data at the edge. The IBM Cloud Satellite allows using cloud capabilities without physically moving data. For instance, when it comes to automobile manufacturing and production line, any amount of latency can hamper the whole production line. But, with 5G, efficient use of transient data is made possible—such that the control plane sits on the cloud and the actual data is at the edge. 

Read: Clash of Tech Titans over Strategic Partnership

Biggest challenges in AI and data analytics 

India has taken a big leap forward in fostering a startup ecosystem leveraging AI and data in the last two years. This can be attested by NASSCOM’s 2022 report, which suggested that AI adoption will add $500 billion to India’s GDP by 2025. However, India’s AI maturity score of 2.45 reveals the untapped potential for AI utilisation in this technological age. 

One of the major roadblocks to widespread AI adoption in this regard has been that 44% of enterprises either have inadequate or siloed data, limiting them from scaling AI solutions. 

There are multiple challenges surrounding this: 

• The whole know-how of capturing the right kind of data, storing data, protecting data, as well as managing data privacy and governance aspects. 
• Making data accessible to all, especially when organisations form integrations with multiple partners.
• Breaking the silos formed by data stored in multiple clouds, along with the organisation’s own on-premise footprint to harness the data across various departments within it optimally.

When asked whether data engineering is a global challenge or specifically pertains to India, Naik stressed that it is absolutely a global challenge. But, when it comes to its adoption, Western Europe is ahead of the curve. And data engineering is particularly important if one needs to infuse AI into their operations enterprise-wide.