loading
loading
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. Inc42: What problem do advertisers still struggle with, and how has Mobavenue built its platform to solve it differently? When did you realise you needed to build more tech in-house? Tejas Rathod: The fundamental problem advertisers face is not simply acquiring users, but acquiring the right users at the right customer acquisition cost, with a clear path to business value. We realised early that existing solutions could help us execute media, but they did not give us the level of control and intelligence required to consistently optimise towards the right customer acquisition cost (CAC). That led us to build our tech stack in-house, layer by layer. Our AI systems analyse billions of signals to identify users with a higher likelihood of conversion, while enabling budgets to be dynamically allocated based on those signals. The objective is not to optimise for media metrics in isolation, but to continuously improve the efficiency and quality of user acquisition. That shift, from media execution to technology-driven outcome optimisation, has been central to how we have built Mobavenue. Inc42: As brands shift from clicks and installs to outcome-based ROI, how does your engineering team build predictability into the platform? What tech decisions have helped Mobavenue operate at the scale of 1.25 Bn data signals a day? Tejas Rathod: Predictability starts with the quality and depth of the signals available to the system. Our engineering stack looks at large volumes of consented data signals and combines them with audience segments, behavioural patterns and relevant user attributes to understand intent. The objective is not simply to identify a user who may click or install, but to predict the probability of completing the business outcome that the advertiser values. For example, in a fintech journey, that could extend beyond an app install to KYC completion, investment or loan conversion. Our Neural Network continuously feeds outcomes back into the system, allowing models to learn which signals correlate with quality conversions and improve subsequent decisions. Adtech is a fast-moving environment where the value of a decision can change within milliseconds, while expectations around accuracy and outcomes continue to increase. So, from the beginning, we made a few deliberate technology choices: real-time decisioning, proprietary intelligence and a common data layer rather than multiple disconnected systems. The other important decision was to build a closed feedback loop into the platform. Every campaign outcome generates a new learning signal that feeds back into our models, helping improve subsequent decisions around audiences, bidding and budget allocation. In FY26, we delivered 42.72 Mn outcomes, and this scale of execution gives our models a continually expanding base of learning. For us, maintaining performance at scale means combining speed with continuous learning, so that every decision becomes more informed by the outcomes that came before it. Inc42: Mobavenue recently formalised its focus on AI infrastructure. How does its Neural Engine think differently from traditional programmatic advertising, and what led you to build an intelligence layer across the marketing lifecycle? Tejas Rathod: Traditional programmatic advertising is largely designed to make the best decision for the immediate media event, such as whether to bid, how much to bid, or which audience to target. Our Neural Engine is designed to learn from what happens after that decision. We have moved our core decisioning from rule-based and traditional ML systems to a unified neural network architecture built around three stages: process signals, predict intent and produce outcomes. It continuously processes signals, predicts intent and evaluates downstream outcomes, creating a closed feedback loop where every outcome becomes a new signal for the next decision. So, intelligence does not simply optimise a campaign; it continuously learns what drives quality outcomes and adapts its decisions accordingly. The adtech ecosystem has become good at automating individual parts of advertising, but intelligence is still fragmented across the marketing lifecycle. Planning, audience selection, campaign execution, creative and reporting often sit across different systems, which means marketers still have to connect the dots and move decisions from one stage to another. We saw that the next opportunity was not simply to automate another task, but to create an intelligence layer that could continuously connect signals, decisions and outcomes. That thinking led to the Mobavenue Neural Engine. It brings together our proprietary advertising intelligence, first-party platform intelligence and real-time business context to make recommendations that become more relevant as the system learns. The larger shift is from AI as an optimisation tool to AI as a decisioning layer across awareness, acquisition and activation. Our Neural Engine is designed to connect planning, execution, optimisation and reporting, while keeping the marketer in control. For us, the future is not just about making advertising faster. It is about making the entire marketing lifecycle more intelligent, connected and outcome-driven. Inc42: The company closed FY26 with ₹218 Cr+ in revenue and roughly a 20.8% EBITDA margin. How closely do you tie your product roadmap to these metrics, and which products drove the growth? Tejas Rathod: For us, the product roadmap is primarily driven by the opportunity we see in the market. Financial performance is an important validation, but we do not build products simply around a revenue or EBITDA target. We look at where consumer behaviour, technology and advertiser demand are moving, and then build around those opportunities. FY26 is a good example of this approach. Our existing platforms, particularly PrsmX and the Neural Engine across all products, were significant contributors to our revenue momentum during the year. These products continued to evolve as we saw greater demand for AI-driven optimisation and outcome-based advertising. PiiX is the next example of that product philosophy. We identified a significant opportunity in the Apple Search Ads ecosystem and launched PiiX in Q1 FY27 as a dedicated intelligence platform for that high-intent environment. So, the roadmap is driven by market opportunity, while revenue and profitability validate whether we are creating meaningful value from those investments. Inc42: Over 73% of your revenue now comes directly from brands. What did you change about the platform and reporting to make enterprise clients comfortable managing campaigns directly? Tejas Rathod: Direct advertiser relationships required us to rethink the platform experience end to end. We undertook a significant overhaul of the interface, underlying infrastructure and reporting capabilities to make the system more robust, intuitive and user-friendly for enterprise teams. We also expanded the platform’s functionality across campaign planning, execution, creative intelligence and measurement. A key focus was making reporting easier to understand while providing the level of granularity that enterprise advertisers need to evaluate performance, understand where their budgets are being deployed and make informed decisions. The objective was to create greater transparency without adding unnecessary complexity. And the larger goal is to make sophisticated technology accessible without requiring the user to deal with the complexity underneath it. Inc42: How has Mobavenue grown over the past year? Can you share any numbers around campaign performance, advertiser ROI or CAC that highlight the platform's impact? Tejas Rathod: Over the past year, our growth has been closely tied to how effectively our technology can improve measurable outcomes for clients. We operate on an outcome-as-a-service model, which means we grow when our clients grow. Revenue per outcome (RPO) is an important business metric for us because it reflects the value generated from every measurable outcome. Our RPO increased from ₹44.99 in Q1 FY26 to ₹48.44 in Q4 FY26, while quarterly outcomes grew from 9 Mn to 12.61 Mn. At the campaign level, this translates into measurable performance. Snabbit achieved 5X growth in installs and 10X growth in first-time purchases through precision programmatic optimisation. For boAt, our PrsmX platform delivered a 91% view-through rate (VTR), along with a 14% increase in aided awareness, a 20% increase in brand favourability and a 19% increase in purchase intent. Inc42: Where do you see the next big opportunities for adtech? Tejas Rathod: I see opportunities across three fundamental areas of brand growth: awareness, acquisition and activation. These are not isolated businesses anymore. A brand may need to create awareness, acquire a user and then activate or retain that user, with intelligence from each stage informing the next. That is the direction in which we are building Mobavenue. Our products address these stages, but they are increasingly connected through a common intelligence layer that allows data and learnings to move across the consumer journey. Inc42: Five years from now, what do you want Mobavenue to be known for? Tejas Rathod: Five years from now, I would like Mobavenue to be recognised as a technology company that covers the broader digital growth ecosystem, not just advertising. We see significant opportunities across adtech, martech and monetisation technology, and our ambition is to build proprietary platforms across these areas that work together intelligently.
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. Inc42: What problem do advertisers still struggle with, and how has Mobavenue built its platform to solve it differently? When did you realise you needed to build more tech in-house? Tejas Rathod: The fundamental problem advertisers face is not simply acquiring users, but acquiring the right users at the right customer acquisition cost, with a clear path to business value. We realised early that existing solutions could help us execute media, but they did not give us the level of control and intelligence required to consistently optimise towards the right customer acquisition cost (CAC). That led us to build our tech stack in-house, layer by layer. Our AI systems analyse billions of signals to identify users with a higher likelihood of conversion, while enabling budgets to be dynamically allocated based on those signals. The objective is not to optimise for media metrics in isolation, but to continuously improve the efficiency and quality of user acquisition. That shift, from media execution to technology-driven outcome optimisation, has been central to how we have built Mobavenue. Inc42: As brands shift from clicks and installs to outcome-based ROI, how does your engineering team build predictability into the platform? What tech decisions have helped Mobavenue operate at the scale of 1.25 Bn data signals a day? Tejas Rathod: Predictability starts with the quality and depth of the signals available to the system. Our engineering stack looks at large volumes of consented data signals and combines them with audience segments, behavioural patterns and relevant user attributes to understand intent. The objective is not simply to identify a user who may click or install, but to predict the probability of completing the business outcome that the advertiser values. For example, in a fintech journey, that could extend beyond an app install to KYC completion, investment or loan conversion. Our Neural Network continuously feeds outcomes back into the system, allowing models to learn which signals correlate with quality conversions and improve subsequent decisions. Adtech is a fast-moving environment where the value of a decision can change within milliseconds, while expectations around accuracy and outcomes continue to increase. So, from the beginning, we made a few deliberate technology choices: real-time decisioning, proprietary intelligence and a common data layer rather than multiple disconnected systems. The other important decision was to build a closed feedback loop into the platform. Every campaign outcome generates a new learning signal that feeds back into our models, helping improve subsequent decisions around audiences, bidding and budget allocation. In FY26, we delivered 42.72 Mn outcomes, and this scale of execution gives our models a continually expanding base of learning. For us, maintaining performance at scale means combining speed with continuous learning, so that every decision becomes more informed by the outcomes that came before it. Inc42: Mobavenue recently formalised its focus on AI infrastructure. How does its Neural Engine think differently from traditional programmatic advertising, and what led you to build an intelligence layer across the marketing lifecycle? Tejas Rathod: Traditional programmatic advertising is largely designed to make the best decision for the immediate media event, such as whether to bid, how much to bid, or which audience to target. Our Neural Engine is designed to learn from what happens after that decision. We have moved our core decisioning from rule-based and traditional ML systems to a unified neural network architecture built around three stages: process signals, predict intent and produce outcomes. It continuously processes signals, predicts intent and evaluates downstream outcomes, creating a closed feedback loop where every outcome becomes a new signal for the next decision. So, intelligence does not simply optimise a campaign; it continuously learns what drives quality outcomes and adapts its decisions accordingly. The adtech ecosystem has become good at automating individual parts of advertising, but intelligence is still fragmented across the marketing lifecycle. Planning, audience selection, campaign execution, creative and reporting often sit across different systems, which means marketers still have to connect the dots and move decisions from one stage to another. We saw that the next opportunity was not simply to automate another task, but to create an intelligence layer that could continuously connect signals, decisions and outcomes. That thinking led to the Mobavenue Neural Engine. It brings together our proprietary advertising intelligence, first-party platform intelligence and real-time business context to make recommendations that become more relevant as the system learns. The larger shift is from AI as an optimisation tool to AI as a decisioning layer across awareness, acquisition and activation. Our Neural Engine is designed to connect planning, execution, optimisation and reporting, while keeping the marketer in control. For us, the future is not just about making advertising faster. It is about making the entire marketing lifecycle more intelligent, connected and outcome-driven. Inc42: The company closed FY26 with ₹218 Cr+ in revenue and roughly a 20.8% EBITDA margin. How closely do you tie your product roadmap to these metrics, and which products drove the growth? Tejas Rathod: For us, the product roadmap is primarily driven by the opportunity we see in the market. Financial performance is an important validation, but we do not build products simply around a revenue or EBITDA target. We look at where consumer behaviour, technology and advertiser demand are moving, and then build around those opportunities. FY26 is a good example of this approach. Our existing platforms, particularly PrsmX and the Neural Engine across all products, were significant contributors to our revenue momentum during the year. These products continued to evolve as we saw greater demand for AI-driven optimisation and outcome-based advertising. PiiX is the next example of that product philosophy. We identified a significant opportunity in the Apple Search Ads ecosystem and launched PiiX in Q1 FY27 as a dedicated intelligence platform for that high-intent environment. So, the roadmap is driven by market opportunity, while revenue and profitability validate whether we are creating meaningful value from those investments. Inc42: Over 73% of your revenue now comes directly from brands. What did you change about the platform and reporting to make enterprise clients comfortable managing campaigns directly? Tejas Rathod: Direct advertiser relationships required us to rethink the platform experience end to end. We undertook a significant overhaul of the interface, underlying infrastructure and reporting capabilities to make the system more robust, intuitive and user-friendly for enterprise teams. We also expanded the platform’s functionality across campaign planning, execution, creative intelligence and measurement. A key focus was making reporting easier to understand while providing the level of granularity that enterprise advertisers need to evaluate performance, understand where their budgets are being deployed and make informed decisions. The objective was to create greater transparency without adding unnecessary complexity. And the larger goal is to make sophisticated technology accessible without requiring the user to deal with the complexity underneath it. Inc42: How has Mobavenue grown over the past year? Can you share any numbers around campaign performance, advertiser ROI or CAC that highlight the platform's impact? Tejas Rathod: Over the past year, our growth has been closely tied to how
document fetched 2026-10-20 02:52 · canonical https://inc42.com/features/mobavenues-tejas-rathod-on-the-shift-from-ad-optimisation-to-ai-decisioning