Navigating the intersection of AI and advertising with Meta's Arun Srinivas

In the realm of digital marketing, Meta has been a prominent player in harnessing advanced technology to reshape advertising strategies. The relationship between Meta and AI dates back to the early days of AI technology. From the inception of AI technologies to their integration into Meta's advertising ecosystem, the trajectory has been one of continuous evolution and innovation.

In conversation with Arun Srinivas, Director & Head - Ads Business (India), Meta, we explore the intersection of AI and Meta, examining how these technologies are reshaping advertising practices. From the inception of AI features to their practical applications in businesses, Srinivas provides insights into the transformative potential of AI in driving business growth.

Edited Excerpts: 

How has the advent of AI impacted Meta?

AI has been the backbone of Meta's operations since 2006, particularly in shaping user experiences and refining advertising strategies. For instance, your feed differs from mine due to AI algorithms that consider factors like interests, followed accounts, and hobbies. This technology enables us to deliver tailored content to users and assists advertisers in targeting the right audience effectively.

Having said that, over the last 18 months, there have been a lot of global conversations and developments around AI. The whole technological landscape has moved up. This period has marked AI's transition into mainstream adoption, accompanied by advancements in innovation, talent acquisition, and technological capabilities, such as computing power. These developments have propelled Meta's commitment to leveraging AI to enhance user experiences and advertising effectiveness.

Can you elaborate on how Meta's generative AI features, particularly in image and text generation, are designed to assist businesses in enhancing their advertising efforts?

AI has played a crucial role in transforming advertising strategies, both on the user side as well as the advertisers side, particularly through features like image and text generation. There’s more than a 20% increase in time spent on our platforms because of these machines which are suggesting content and eventually reading to higher engagement. 

All these pros have now gone into what we just launched which will help businesses find their audiences on our platforms. For example, businesses can now utilise AI to optimise their ad creatives, ensuring relevance and effectiveness across various platforms. Whether it's crafting compelling captions or generating visually appealing images, AI empowers businesses to enhance their advertising efforts and engage with their target audience more effectively.

Could you share any insights or data regarding how these AI-driven features have positively impacted ad performance or engagement rates during the testing phase?

Over the last 18 months, we have launched a suite of AI products for all our advertisers — larger and small. It's called the Advantage Plus Suite. 

The implementation of AI-driven features has yielded promising results in terms of ad performance and engagement rates. For instance, app install campaigns have seen a notable decrease in costs, with Dream11 experiencing a 12% reduction in expenses. 

While performance marketing has taken off over the last couple of decades, a lot of it is still left to humans. So what we have done is actually allowed the engine to minimize the variable sought for a campaign, where they can just define the overall budget. By leveraging AI to streamline campaign management and optimise audience targeting, businesses can achieve better outcomes and maximise their advertising ROI.

As Meta aims to make these features globally available, how do you plan to address localisation and cultural nuances to ensure the generated content remains relevant and effective across different regions and markets?

Our vision for AI extends beyond individual markets, aiming to cater to diverse cultural contexts and linguistic nuances worldwide. While achieving 100% localisation may pose challenges due to the complexity of certain regions, we are committed to continuously improving our AI models to accommodate diverse cultural preferences and linguistic variations. Through ongoing refinement and expansion of our language models, we strive to make AI-driven content more relevant and accessible to users across the globe.

Can you explain how centralising the generative AI features under Advantage+ Creative streamlines the ad creation process for businesses? What specific advantages does this integration offer compared to using individual tools separately?

Centralising generative AI features under Advantage Plus Creative streamlines the ad creation process by providing businesses with powerful tools to optimise their advertising content. With features like image and text generation, businesses can create tailored ads more efficiently, allowing for greater customisation and flexibility. This centralised approach simplifies the ad creation process, empowering businesses of all sizes to produce compelling and relevant content that resonates with their target audience.

For large businesses, a big challenge was that we were not able to customise and personalise to each market and cohort. Now with the feature in place, images can be generated dynamically making it easier for businesses of all sizes. 

Beyond just facilitating ad creation, how does Meta support businesses in analyzing and optimizing their campaigns using AI-driven insights or performance metrics?

We already have a dashboard and tools which every advertiser can leverage for every campaign. The image generation will work in conjunction with the Advantage Plus feed and equally could work as a silo as well.

One advantage is the creative function, where AI can choose which creator to serve to whom, ensuring better efficiency and personalisation for higher ROI. Secondly, we also encourage the people who are in the learn and test phase. For example, for a shampoo brand, you can create multiple creators around the same product and serve the right creative to the right person.

What additional enhancements or features can businesses expect from Meta's generative AI capabilities in the near future? Are there any specific areas of focus or industry trends driving these developments?

We see a lot of Gen AI that we are already leveraging through the messaging ecosystem. We equally see a lot of Gen AI in chatbots and enhancing conversational experiences. These developments are driven by the increasing adoption of AI in various industries, particularly in improving customer interactions and efficiency.

Lastly, as AI continues to play a significant role in marketing, how does Meta ensure ethical use and transparency in the deployment of these AI-driven features, particularly in areas like data privacy and algorithm bias?

We take privacy very seriously. A lot of businesses own it, but Meta can't trade this information in any form. There are enough technologies in terms of conversions API for which the signal is encrypted and the data is private.     

Privacy is paramount, with encrypted data and transparent practices. We're committed to responsibility, ensuring AI recognizes diversity and avoids bias. Tools are also used to remove harmful content, reflecting our commitment to fair and ethical practices.

Rapid Fire

Can you describe Meta's Gen AI capabilities for businesses' ads, covering all its features?

Meta's Gen AI features include image generation, text generation, and image enhancement. Advertisers can create various image variations, and catchy slogans, and enhance image quality, empowering businesses to generate ads quickly and effectively.

Could you share some examples of where these features fit best?

Small businesses, like boutique coffee plantations, can benefit by quickly generating images and catchy slogans for their ads. For example, they can generate images of people enjoying coffee or create captivating captions using keywords related to their products.

Can you share some tips for businesses that would use the feature?

I recommend businesses experiment with the features to understand what works best for them. AI is still evolving, so experimentation will help refine outputs and improve effectiveness collectively.