The initial frenzy around generative AI gave the marketing industry an unprecedented engine for content creation. Yet, as digital channels drown in automated copy, generic social posts and creative clutter, the fundamental challenge for brands is no longer about producing more; it is about knowing what actually drives business outcomes.
At CMOs’ Charcha – Bengaluru Chapter 2026, a fireside chat titled ‘GenAI to Predictive AI’ tackled this pivotal transition. Moderated by Rajesh Gouri, Lead – AI, CereOne Mediathe panel brought together Praseed Prasad, President – Growth & Marketing, South Asia, WPP Mediaand Sooraj Balakrishnan, Head of Marketing, Acerto unpack how brands are moving past the hype to build connected, predictive data systems without losing human ingenuity.
Nobody is asking for more AI anymore
Praseed Prasad opened the discussions with something that has quietly become true across the industry. Clients have stopped walking into meetings asking how they can use more AI. They are walking in with the same old problems, market share, retention, lifetime value, and AI is simply one of the tools now on the table to solve them.
What has genuinely changed underneath all of it is consumer behaviour itself. The old idea of a predictable monthly purchase cycle has broken apart into a scattered mix of general trade, modern trade, quick commerce and marketplace behaviour, with impulse and bulk purchases sitting side by side.
Prasad’s team spends most of its energy reading these signals through data partnerships before AI even enters the conversation, building consumer cohorts that get addressed differently instead of the old one message for everyone approach.
Knowing what not to make
Sooraj Balakrishnan brought the conversation to a place most AI panels avoid. He said the real tension at Acer right now is not quantity versus quality. It is figuring out what should never be published in the first place.
That is a strikingly different way to think about a generative tool. Most brands measure AI by how much it can produce. Balakrishnan measures it partly by what it helps them filter out before it ever reaches an audience. He was blunt about the risk on the other side too, calling out what the industry now casually refers to as AI slop, content that exists purely because it was cheap to generate.
Where AI earns its place at Acer is prediction rather than production. With close to 10 million PCs already in its install base the brand is using purchase history, warranty registrations and browsing signals to figure out who is close to upgrading, who is likely to cross purchase and who might respond to a specific warranty offer.
The goal, Balakrishnan said, is not just knowing the customer. It is predicting what that customer is going to need before they ask for it.
None of that works without clean data underneath it. Acer built a data lake specifically to stop different channels from handing over information in incompatible formats, because as Balakrishnan put it plainly, garbage in still means garbage out no matter how advanced the model on top of it is.
Prasad described a similar problem being solved differently at agency scale. WPP Media built what it calls WPP Open, an operating system that behaves like a federated data clean room. Client data never actually leaves its own servers. It simply talks to other data sources such as telecom or marketplace data inside private workspaces that keep every client’s information walled off from every other client’s. The approach puts privacy into the architecture itself, allowing different data sources to work together while keeping client data separated.
The Ben Affleck line that reframed the whole panel
The most memorable moment of the session had nothing to do with dashboards. Prasad brought up an interview he had watched over the weekend where Hollywood actor Ben Affleck was asked whether AI could compress a show like House of Cards down to two or three seasons in a single year. Affleck’s answer was that AI behaves like a skilled craftsman who knows exactly how to execute a task. But art and the artist are defined by something else entirely, knowing when to stop.
That line became the quiet spine of the entire second half of the discussion. Balakrishnan echoed the same instinct from the brand side, saying AI can crunch complexity and surface predictions, but has no real feel for the cultural and linguistic nuance a country as varied as India demands from a brand’s identity. Prasad added a warning aimed squarely at newer talent in the industry, describing how easy it has become for a junior team member to feed a brief into an LLM and present a finished deck without ever having genuinely worked through the thinking themselves.
Both speakers landed on the same underlying principle even though they arrived from different directions. AI is excellent at anything objective. The moment a decision becomes subjective, brand voice, tone, cultural read, it needs a human sitting in that seat and not just supervising the output.
The metrics marketers may need to rethink
The panel closed on a question that exposed just how far this thinking has traveled beyond content creation and into measurement itself. What is one metric that marketers should stop tracking.
Balakrishnan said impressions, arguing that the traditional linear funnel simply does not describe how people buy anymore. A customer today might discover a product through an Amazon review, verify it on an unrelated site and convert somewhere else entirely off the back of a completely different offer. Attributing that journey to a single impression is measuring the wrong thing entirely.
Rajesh Gouri went further and named search volume itself as the metric he would retire, arguing the entire discipline has shifted from keyword guessing to people expressing full intent in their own words across different phrasings and languages that all mean the same thing.
Prasad pushed back gently on both, arguing nothing needs to be dropped outright so much as reinterpreted. Even impressions still carry meaning if you can finally attribute them properly instead of watching them vanish into ad fraud and unclear measurement the way the industry has for years.
Gouri closed the session with the line that tied the entire hour together. GenAI gave marketers a voice and a megaphone. Predictive AI is giving them a radar and the discipline to know what to do, where to stop and what is likely to work before a single dollar gets spent.
