Taming the data deluge with AI

As businesses face demands for quicker decision-making, traditional analytics methods are becoming inadequate — especially amid rapid data growth. In this Q&A, two experts from Wipro –– Srinivasaa “Srini” HG, senior VP & global head of data, analytics & AI, and Balasubramani “Bala” H, GM & global head of data transformation –– explain how agentic artificial intelligence technology combines with well-structured data products to provide a solution, enabling organizations to generate real-time insights and automate business processes.

Q: Why are traditional data analytics methodologies and strategies no longer adequate?

Bala: Traditional data analytics was based on predefined rules and consumed after the fact. Today, we need output right here and now. Customer requirements change on a real-time basis, so we need to make sure answers come in real time. And business users are becoming citizen data scientists, with the capability to build their own analytics, create their own assessments, and do so on the fly.

Srini: With the performance and the kind of data that a data intelligence platform brings together, people can create new reports to answer deterministic questions very quickly. With the emergence of LLMs, users don’t want to just view reports; they want to interact with them. They want to create their own reports on the fly.

Q: What are some of the key elements of a data strategy that enable companies to make more effective use of their data?

Bala: We conducted interviews with more than 150 industry chief data & analytics officers to understand the challenges they face in the new age of data. Five key things came out. First, companies don’t effectively use theirinternal, proprietary data. Data internal to the org is the most important asset to ensure business outcomes for your enterprise.  Second is democratization across the data estate, which makes it easy for everyone to access and effectively use the data they need, along with the tools to do it.

Third, things are changing rapidly, and there is a big gap between how quickly companies can adopt new technologies and what is available in the market, from AI to genAI and now agentic AI.

Fourth is data governance, which has too often been an afterthought. With the advent of AI, governance needs to be mainstream. For AI to be successful, you have to have the right data and governance as your foundation and ensure data is properly managed (has lineage, quality, observability, and is only available to those authorized to see it).

And the fifth key element is the talent deficit in the market to build at scale, along with the fact that many systems will be human-plus-agent-driven (“Hugentic”), not purely human. So, you have to adapt to that.

Srini: Metadata, lineage, ontology, and taxonomy of data are all becoming critical. A definition of a field in one table or database may be radically different from another, even though the name of that field is the same. An expiry date could refer to a product or a contract. Today’s models can deliver at best 80% or 90% accuracy because they aren’t trained on the data, definitions, meaning, and context of a particular enterprise, which is not good enough. What we see as an architectural block is fine-tuning and training models for enterprise adoption. 

Q:  What’s the vision behind the idea of composable data elements?

Bala: Business requirements and questions are changing frequently. Having things pre-built –– dashboards, aggregated data sets, and so on –– is not going to help as we move forward. Composability introduces a plug-and-play feature, allowing different layers that are part of your data life cycle to be easily separated, changed, replaced, or enhanced without impacting the others.

Srini: These composable data products could also be created based on the queries or data that an agent determines it needs in order to meet the end user’s needs. Managing that element is a much larger challenge to solve from an architectural standpoint. The moment you talk about AI, guardrails become important for dealing with explainability, responsibility, hallucination, jailbreaks, and so on. And those also need to be composable in nature, like a super agent managing another agent.

Q: What is your company doing to help companies prepare for this new data environment?

Srini: We have two strategic AI delivery platforms, WEGA and WINGS, as part of Wipro Intelligence™, which is our unified suite of AI-powered platforms and solutions. Both are strengthened by partner technologies, including those provided by our partner, Databricks. Building a new application or data pipeline, training an existing model, or creating a domain-oriented small-language model –– these are all build activities enabled by WEGA. WINGS covers all capabilities necessary to manage and run applications, infrastructure, and business processes securely.

Bala: We are a Databricks Global Elite Partner, which brings advantages to what we can do for our customers. We have more than 200 use cases implemented across 80+ customers in all key industries. For example, WealthAI is a solution on which we are working closely with Databricks for our banking and financial services customers, to help them deliver the right solution based on the business value it generates. WEGA and WINGS agents and AI assistants are also compatible with and complementary to Databricks.

To learn more, visit us here.