Internet has moved from words to video: Convosight’s Tamanna Dhamija

The fourth edition of MarTechAI Summit, held on December 10, set the stage for the exploration of how data, automation and generative AI are redefining the future of marketing. Among the powerhouse speakers was Tamanna Dhamija, CEO & Founder of Convosight Analytics Pvt Ltd, who delivered the “Convosight Spotlight Session”. 

Her session, titled “Attention, Emotion & Influence: How Convotrack & GenAI Is Rewriting Consumer Understanding Through Video Intelligence,” focussed on the dramatic shifts in how consumers behave online and why brands today need a next-generation intelligence layer to keep up.

Dhamija began with an analogy, recalling how 10 years ago travellers relied on printed maps, whereas today they depend on Google Maps, ChatGPT and Perplexity, constantly evolving, real-time ecosystems. She said consumer understanding needs a similar transition from static insights to dynamic, always-on intelligence. 

She highlighted how India’s digital population of nearly 800 million now spends close to five hours online every day, and 80% of that time is on social platforms where people are consuming and creating video-first content. She noted that nearly 70% of users are active on at least one major platform such as Instagram, YouTube or Facebook, and that platforms like TikTok generate roughly 4 trillion posts every month, while YouTube alone sees 500 hours of content uploaded every single minute. Emphasising the magnitude of this shift, she said consumers are already answering every business question through this content explosion; what brands lack is the capability to bring those answers alive.

This context set the stage for why Convosight built Convotrack, its GenAI-powered social and video intelligence engine designed to decode real-time behaviour from the environments where consumers actually express themselves. 

A major portion of her session centred on how Convotrack fundamentally differs from tools like ChatGPT, Perplexity or generic large language models. Explaining the gap, she said, “LLMs look at text. At best, metadata. But consumer expression today is video, visuals, voice, tone, pauses, chaos, context. That entire layer is lost.” 

She elaborated that most LLMs rely heavily on open-web text, while real consumer conversations do not exist on longform blogs or static articles anymore. “Consumers aren’t writing blogs. They’re on Instagram, YouTube, Reddit, X, creating raw, unfiltered content,” she said. Convotrack pulls directly from these social environments and overlays them with comments, resonance patterns and behavioural signals to construct a more complete, real-time reflection of how people think and behave.

Dhamija underlined that Convotrack is not only trained on social behaviour but on brand minds, thousands of briefs, insights frameworks and research structures accumulated from CMOs, brand heads and consumer insights teams. This makes the platform operate almost like a decision-making brain, enabling functions from segmentation and content resonance to competitive intelligence and category code mapping at what she called “the speed of curiosity.”

During the session, she showcased multiple videos, demonstrated how Convotrack interprets emotional intent and narrative patterns, and engaged the audience through a series of interactive Q&As. In one of the video examples, she illustrated how Convotrack scanned more than 2,600 posts, analysed nearly 84,000 conversations, decoded patterns across 320 million video views and reviewed close to 22 million engagements, helping the brand uncover not just what people were saying about protein but also the cultural cues, emotional triggers, misconceptions and emerging motivations shaping the category.