Mobavenue’s Tejas Rathod On The Shift From Ad Optimisation To AI Decisioning

by · Inc42

SUMMARY

  • As digital advertising shifts from clicks and installs towards measurable business outcomes, AI is increasingly being used to predict user intent and optimise decisions beyond the immediate media event
  • In a conversation with Inc42, Mobavenue’s founder and CTO Tejas Rathod explains how the company is building AI-driven advertising infrastructure around real-time decisioning, privacy-conscious signals and continuous learning
  • From processing more than 1.25 Bn data signals a day to connecting awareness, acquisition and activation, Rathod discusses how he sees intelligence evolving across the digital marketing lifecycle
  • Added to Saved Stories in Login

A high click rate cannot tell a fintech company whether a user will complete the KYC. Similarly, downloading an app does not promise a purchase. Yet these remain the metrics that digital advertising has traditionally been built around.

For brands, the real challenge is understanding whether a click or app download will actually translate into a business outcome. As advertisers demand clearer links between media spending and business outcomes, adtech is moving towards systems that can predict intent, learn from downstream behaviour and optimise for what happens beyond the click or install.

This is the shift Mobavenue is targeting. The company processes more than 1.25 Bn data signals a day and has built its technology stack around real-time decisioning and continuous feedback from campaign outcomes. In FY26, it delivered 42.72 Mn outcomes, giving its models a growing base of data to learn from.

At the centre of this is Mobavenue’s Neural Engine, which connects signals, intent and outcomes. Rather than treating a bid, click or install as the endpoint, the system is designed to learn from what happens after that decision and use those outcomes to inform future audience, bidding and budget decisions.

For Tejas Rathod, founder and CTO of Mobavenue, this marks a broader shift in advertising — from AI as an optimisation tool to AI as a decisioning layer across the marketing lifecycle.

In a conversation with Inc42, Rathod explains the technology behind Mobavenue’s platform, how the company approaches privacy and invalid traffic, and why it is building towards an advertising ecosystem that connects awareness, acquisition and activation. 

Here are the edited excerpts…

Inc42: What has been the biggest shift in the Digital advertising industry, and what is driving it?

Tejas Rathod: The biggest shift is the movement from automation towards predictive intelligence and AI autonomy. Earlier, technology helped advertisers execute decisions faster. Today, AI is increasingly being used to anticipate intent, learn from outcomes and make decisions dynamically.

This is being driven by the growing complexity of digital ecosystems, the volume of signals available and the pressure to deliver measurable business outcomes. The next phase is moving from AI as an optimisation tool to AI as an increasingly autonomous decision-making layer.

Inc42: With privacy regulations tightening and invalid traffic remaining a challenge, how has Mobavenue prepared its tech to maintain targeting accuracy and campaign efficiency? 

Tejas Rathod: Privacy has been a foundational consideration in how we have built our tech. Our platforms operate within the requirements of applicable privacy frameworks and regulations across the markets we serve, including the General Data Protection Regulation (GDPR), California Consumer Privacy Act (CCPA) and India’s Digital Personal Data Protection (DPDP) framework. We also use privacy-conscious approaches to identity, including ID masking and hashing, to derive intelligence without unnecessarily exposing identifiable information.

Our models rely on aggregated, consented signals and behavioural patterns. This lets the system understand user intent and quality while maintaining data safeguards. The objective is to make privacy and targeting intelligence work together, rather than treating them as competing priorities.

Invalid traffic needs to be addressed at multiple points in the advertising lifecycle. Filtering it only after a campaign has run can still result in wasted spend. We therefore use internal invalid traffic detection capabilities at both the pre-bid and post-bid stages.

At the pre-bid stage, the system evaluates available signals and supply characteristics before deciding whether to participate in an auction. This helps reduce exposure to questionable inventory. Post-bid, we analyse delivery and campaign signals to identify patterns that may indicate invalid activity and use these learnings to improve subsequent optimisation.

Fraud detection is also part of our broader machine learning capabilities, alongside audience segmentation, bid optimisation and real-time learning. The objective is to protect both media efficiency and the quality of outcomes.