“Information governance is the same as belief. Can I belief my information to run AI fashions on it? Can I belief my information to make choices primarily based on it?” stated Varun Babbar, vice chairman and India managing director at Qlik, summing up the core theme of Qlik Join 2026 – a shift from experimenting with AI to deploying it at scale, rooted in belief in information.
In dialog with indianexpress.com, Babbar shared that when he meets enterprises immediately, he finds most AI conversations fall into three classes. The primary is democratisation, which is making AI accessible to workers who aren’t information scientists. Second is about conversational analytics, basically permitting customers to question enterprise information the best way they already work together with client AI instruments at house. And the third, essentially the most distinguished, is agentic AI – methods that don’t merely output an perception but additionally act on it.
“At house we’re utilizing ChatGPT, Claude, and Gemini for nearly every little thing. Gone are the times of utilizing the search bar. Individuals now need the identical conversational expertise with their enterprise information – they wish to ask a query and get an perception that helps them take motion,” Babbar stated.
No matter whichever of those three matters a dialog begins with, Babbar identified that it nearly all the time circles again to the identical underlying concern – information. “You can not realise any of those use instances with out the fitting information basis,” he stated. And this attracts enterprises into deeper discussions about governance, high quality, lineage, and structure. Although much less glamorous, they’ve emerged as extra consequential work that determines whether or not an AI funding by enterprise pays off.
India as an early adopter
When requested about India’s enterprise AI market, Babbar gave a extra nuanced take. In keeping with him, India is an early adopter when it comes to efficiency-focused purposes. This implies firms are centered on deployments the place AI improves particular person productiveness as a substitute of reworking complete enterprise processes. “The change that has occurred is that over the past yr or two, individuals had been throwing each drawback on the LLM. Most had been failing as a result of not each drawback is to be solved by a single answer,” he defined.
“With time, individuals have realised the significance of the fitting information basis, and now organisations are driving real worth from these investments.”
This evolution displays a world sample. Within the early years of the present AI wave, enterprise pilots unfold quickly whereas manufacturing deployments lagged. The proof-of-concept section generated enthusiasm however was restricted when it got here to outcomes. Babbar argued that the method itself was the issue. “Earlier, individuals had been extra inclined in direction of experimenting as a result of it was thrilling. Organisations had been throwing every little thing at AI – a whole lot of experiments, a whole lot of proofs of ideas – however they weren’t getting productionised as a result of there was no actual worth behind these use instances.”
Story continues under this advert
Nevertheless, immediately, he shared that organisations are ranging from the result and dealing backwards to the know-how. It is a reversal, in keeping with him, that has meaningfully improved the speed of profitable deployments.
The info conundrum
Through the dialog, one of many extra hanging observations Babbar supplied involved the composition of enterprise information. In keeping with him, in most organisations, greater than 80 per cent of information is unstructured, that means it consists primarily of emails, PDFs, and paperwork, whereas databases account for under round 20 per cent. Nevertheless, analytical efforts have traditionally centered nearly completely on that 20 per cent.
“There’s a whole lot of worth sitting in the remainder of the 80 per cent. Organisations at the moment are revisiting their architectures on how they’re going to combine structured and unstructured information and analyse them collectively,” he defined, including that that is the issue Qlik Solutions was constructed to deal with. The device, first launched to work with unstructured information, has since been prolonged to assist simultaneous queries throughout each structured and unstructured sources. Babbar additionally pointed to the corporate’s information merchandise idea, which is a ruled, searchable market the place customers can discover and belief enterprise information units, just like how customers search an e-commerce platform for a product of their alternative.
In keeping with Babbar, this urgency of getting information governance proper has been additional sharpened by regulatory developments. India’s Digital Private Information Safety Act, 2023, introduces penalties for mismanagement, including a compliance dimension to what was beforehand a purely strategic dialog. Even within the public sector, Babbar has noticed a notable shift. “During the last yr or two, I’ve seen many RFPs coming even from public sector entities speaking about consolidating information and creating golden information for residents in order that they’ll have a ruled, reliable mechanism for his or her companies.”
On the transfer in direction of Agentic AI
Story continues under this advert
One of many key highlights of Qlik Join 2026 was its push into agentic AI, which was exemplified by its providing of Uncover Agent, which is a system that has been designed to supply insights from information repeatedly with out human prompting. Extra distinguished than any single product, nonetheless, is the underlying imaginative and prescient, which is brokers from a number of platforms working collectively to shut the loop between perception and motion.
Babbar demonstrated this level with a sensible state of affairs. Below present circumstances, a person spots a list shortfall on a dashboard and alerts a colleague, who raises a ticket in a system like ServiceNow, and everybody tracks decision via e mail. In an agentic structure, the identical chain of detect, resolve, and act occurs mechanically. “As quickly as an agent finds an perception and determines there may be going to be a scarcity of a specific stock at a particular retailer, a ticket is mechanically raised within the procurement system. No guide intervention. The motion is delivered at a a lot quicker tempo.”
The Qlik government shared that that is additionally the logic behind the corporate’s just lately introduced partnership with ServiceNow. That is additionally why he’s bullish on the Mannequin Context Protocol, an open customary that enables brokers constructed on totally different platforms to speak with one another. “The extra you collaborate, the extra the platform turns into open. Constructing a extra open ecosystem and extra collaborations inside organisations goes to be crucial to creating this work at scale.”
Governing the tempo of change
For all its promise, Babbar is candid in regards to the governance problem. The tempo of AI growth, new fashions arriving nearly weekly, creates stress on firms to maintain rebuilding. Nevertheless, his commentary is that the organisations navigating this in one of the simplest ways are those who have formalised their AI decision-making. “Firms have constructed AI governing councils. There are Chief Information Officers as a definite designation now and Chief AI Officers in a number of the bigger organisations. They’ve constructed a mechanism to control their AI tasks, which was not there a number of years again. Now it’s run as a extra ruled, structured programme – and therefore the know-how turns into nearly interchangeable.”
Story continues under this advert
The implication right here is that the organisations extra prone to profit from AI aren’t these chasing the newest fashions however these with the information foundations, governance buildings, and use-case self-discipline to deploy it meaningfully. On that measure, Babbar urged, the dialog in India has already moved a good distance.