AI adoption accelerates when proper data plan is in place: Wipro CTO

Wipro has been doing AI for a long time but with the acceleration of generative artificial intelligence (Gen AI), says Arun

Wipro Chief Technology Officer Sandhya Arun is readying the IT services firm to work in frontier technologies such as quantum, agentic artificial intelligence (AI), and blockchain, as she believes these will converge with AI in the not so distant future. In a conversation with Avik Das in Bengaluru, Arun talks about Wipro Innovation Network, importance of data and what are some of the key pillars enterprises should build to become AI-powered. Edited excerpts: 

How do you think Wipro Innovation Network will help the company move ahead in its AI-first approach? 

Our priority is an AI-powered Wipro strategy, where we talk about 'deliver-better and operate-better'. If you look at the Wipro Innovation Network, while AI is dominant and pervasive and there is an evolution of innovation quotient, we also have other frontier technologies we want to tap and be ready such as quantum and edge computing. We want to focus on five strategic technology frontiers going forward: agentic enterprise which is how enterprises will look when agents are fully mature, robotics with embedded AI when robots start getting intelligent, distributed ledger in blockchain technology which is making a comeback in certain countries as some European governments have started sponsoring, and cyber resilience in a quantum and AI safe space.  As these technologies mature, the solutions will be combinatorial. So we need not be just AI-powered but also ready for frontier technologies when they converge with AI. 

What are some of the pillars for your AI strategy? 

Wipro has been doing AI for a long time but with the acceleration of generative artificial intelligence (GenAI), we have reconfigured our priorities to make sure what we do and how we run the company is AI-powered. Whether it is CEO and CFO operations to delivery, we look if it has some AI solutions that make sense. It could be simple machine learning (ML) operations, automations, or could be AI agents. Every service line is focused on transforming how they offer solutions to clients and all clients are focused on how to adopt AI in a responsible manner. So data governance, responsible AI are some of the key pillars. Besides this, culture is also important. Are you hierarchical, risk averse, do you know how to manage risk and innovate at the same time. Some of these are important to be an AI-powered enterprise. 

Are most of these Gen AI projects still at a proof of concept (PoC) stage or moving into production to deliver value? 

We are seeing that shift happen which is significant in terms of conversion. All sectors which are moving fast such healthcare, BFSI, manufacturing, and energy have seen this change. 

GenAI adoption has been slow and low across enterprises, according to analysts and consultants, in spite of the technology being present for more than two years. What are the reasons for it? 

One of the reasons is that evolution was faster than we expected. And I think to some extent, business leaders started putting pressure on the technology to show them some business value. Business value occurs when you have a perfect foundation for AI adoption. And AI adoption can happen when your data strategy is in place and data is organised. If not, then AI will be as good or as bad as the data because the intelligence comes from it. Data governance and strategy are huge headwinds in terms of accelerated adoption. The next impediment is integration. More and more organisations are looking at HR, business and IT coming together to get ready for the agentic world as it impacts people, productivity, and business. 

So does that mean enterprises have to get their data in a more structured format to derive the benefits of AI? 

The beauty is not turning the unstructured data into a structured one. The need for structured data is not as high as it used to be in the past. Previously it was essential for information systems and databases to have data in a structured manner. With AI, even if you feed unstructured data or have a source of unstructured data, AI can still make sense of it. The only thing is when you are designing, you should be able to architect and orchestrate it with humans in the loop to validate and get the feedback. 

Unstructured data can have more results in human engagement spaces such as employee satisfaction and customer satisfaction.