How will generative AI impact programmatic media buying?

Generative AI has made significant strides in enhancing programmatic advertising by enabling hyper-personalisation and optimising ad creatives, placements, and bidding strategies.

Ad giants like Meta, Microsoft, Amazon, and Google are rapidly advancing their machine-learning capabilities to enhance their AI-generated programmatic advertising features.

Meta introduced new machine learning tools and an AI Sandbox for experimental generative AI advertising, and Amazon upgraded its systemwide machine learning and predictive algorithms.

Meanwhile, Google announced AI-powered ad products, including generative AI in Performance Max, enabling advertisers to develop creative assets and headlines through natural language conversations with AI.

Google plans to show AI-generated ad content across its platforms, testing AI-based search experiences like ‘AI Snapshot’ and providing AI-generated summaries and sponsored links in search results.

Campaign looks at how the industry is approaching AI-generated programmatic advertising.

The agency perspective

Media investment holding group GroupM predicts that by 2032, 90% of media will touch AI in one way or another.

Its performance media arm GroupM Nexus, which unites GroupM Services, Xaxis, Finecast, and AI technology Copilot, has embraced generative AI to create more relevant content, target specific audience segments, and predict ad performance for improved return on ad spend (ROAS).

Deepika Nikhilender, executive director of programmatic, Asia Pacific at GroupM Nexus, believes that as connected TV (CTVs) and streaming services become dominant, generative AI can play a pivotal role in contextual targeting and dynamic ad insertion.

One of the ways it could play a role is in contextual targeting. Generative AI models can be used to analyse the context and content of CTVs and then generate insights on the most suitable ad placements based on the data obtained.

Using generative AI, advertisers can target their ads towards specific content categories per their target audience's interests and preferences by understanding various shows or events' themes, moods, and genres.

(L-R) Rahul Vasudev (Scibids), Deepika Nikhilender (GroupM Nexus), Tom Jones-Barlow (SeenThis), Matt Tindale (LinkedIn), Nick Seckold (Microsoft Advertising), Mitch Waters (The Trade Desk)

“This ensures that ad placements are relevant and practical, improving advertising outcomes. Additionally, generative AI can help with Dynamic Ad Insertion. Generative AI enables real-time insertion of personalised and targeted ads into CTV streams by analysing viewer data,” Nikhilender explains to Campaign.

“Through this, generative AI models can determine the most suitable ads to be inserted, ensuring viewers receive relevant and current advertisements. This helps advertisers deliver personalised messages at scale and in a timely fashion.”

However, Nikhilender notes that challenges like data quality, privacy compliance, and user experience must be addressed when applying generative AI to new channels.

“The use of generative AI has raised many concerns—namely, compliance and ethical issues. Compliance with privacy regulations and ethical guidelines is paramount when utilising generative AI in new channels or formats,” she explains.

“Advertisers and marketers must respect user privacy preferences and abide by relevant data protection regulations while generating personalised ads.”

On the positive side, Nikhilender says generative AI can help ensure brand safety by verifying content, monitoring brand reputation, and detecting and preventing ad fraud. 

For example, the AI models can undergo training to identify patterns that suggest ad fraud. By analysing extensive datasets and real-time ad performance data, AI algorithms can detect irregularities, alert brands about suspicious activities, and protect their advertising campaigns against fraud.

“Generative AI can support brand monitoring by analysing text, sentiment, and context across various platforms like social media, websites, and digital ads. It identifies situations where a brand's reputation is vulnerable to misleading or contains fraudulent content,” explains Nikhilender.

“Generative AI also provides real-time alerts when brand safety risks or fraudulent activities are identified during ad campaigns. This empowers brands to respond quickly and take necessary measures to safeguard their reputation.”

The ad platform perspective

Microsoft Advertising, which has evolved from Bing Ads, is looking to use the rise of generative AI to support brand marketers in Asia.

Microsoft Advertising’s ecosystem combines the Bing search engine, Microsoft news distribution, the Windows operating system, LinkedIn, the Edge browser, and the gaming platform Xbox. Investments in AI, partnerships with OpenAI, and representing Netflix's ad tier have further enhanced the ad platform’s offerings.

For example, Microsoft has been experimenting with ads in Bing Chat responses since February. During the pilot phase, users may encounter ads, but Microsoft has not disclosed the exact number of users seeing them. 

Nick Seckold, the regional vice president of advertising sales for Asia at Microsoft, notes that the integration of ChatGPT in Bing search has led to significant growth, surpassing 100 million daily active users and accumulating half a billion chats in three months.

“Being embedded into the Edge browser and powering other publisher platforms, Bing's quality and impact reach further than some may realise,” Seckold tells Campaign.

“This integration has provided users quicker access to the information they seek, resulting in more substantial ROI and increased satisfaction for consumers and advertisers alike. While it is still early to present case studies, the initial results have shown promising signs of success, with users finding what they need more efficiently.”

According to Seckold, the Bing chat acts as a co-pilot, curating results, suggesting answers, and refining searches, ultimately reducing the need for users to click through multiple links to find relevant information.

This co-pilot concept extends beyond Bing Chat. Microsoft incorporates it into various products and services, aiming to simplify and enhance the user experience throughout the Microsoft 365 family of apps and services.

However, the new features do not come cheap. Microsoft has said it will charge at least 53% more to access co-pilot, which will cost $30 in Microsoft 365. The additional cost is voluntary and adds to existing monthly plans, potentially tripling expenses for some Microsoft customers, with plans ranging from $12.50 to $57 per user per month.

“AI empowers media buyers and campaign managers by enabling them to prompt the AI to handle campaign setup, optimisation, reporting, and insights generation,” explains Seckold.

“Instead of manually clicking through various parameters for campaign creation, AI allows users to instruct it to set up a campaign with specific targeting, audience demographics, and budget.”

Seckold points out generative AI's capabilities extend to creative aspects, such as generating images for ads. In the social space, where carousel ads may require multiple creative types, advertisers can prompt the AI to create various creators based on specific specifications and choose the most suitable ones.

“Though these functionalities are not widely available yet, understanding the current AI capabilities and their potential suggests that campaign buyers, planners, and managers will soon be able to utilise AI to efficiently prompt, set up, run, optimise, and analyse campaigns,” explains Seckold.

“This integration of AI into campaign management processes presents an exciting prospect for streamlining and enhancing advertising efforts.”

The media buying platform perspective

To ensure the ethical and transparent use of consumer data in programmatic advertising, Scibids uses AI technology that does not rely on analysing navigation history tracked through cookies or personal identifiers.

Instead, it uses non-user-specific metadata from bid requests to create better alignment between brands and consumers, maintaining consumer privacy.

Rahul Vasudev, the managing director for APAC at Scibids, points out that generative AI generates new content based on patterns learned from previous data, allowing for custom bidding scripts and algorithms.

For example, Scibids worked with a home improvement retailer that defined its ad effectiveness as return on ad spend (ROAS), measured by Google’s DoubleClick Campaign Manager. After switching to a custom algorithm, the results were a 27% increase in ROAS, incremental revenue upwards of $2.4 million, and a CPM efficiency of 31%. This translated into effectively an ROI of 59 for every dollar paid for the AI tech.

“Custom algorithms let brands go deeper, using the full range of contextual signals and putting a business-specific lens on media buying. Digital media presents exponential cardinality for optimisation,” Vasudev explains to Campaign.

“Just optimising on ZIP codes can incorporate tens of thousands of rules, and humans cannot upload those rules daily or weekly. Custom AI harnesses the granularity of the modern digital media landscape to fuel significant cost compression. Custom algorithms can double clients’ media productivity overnight.”

Over at The Trade Desk, the platform has introduced Kokai, which incorporates distributed AI, deep learning algorithms, and advanced measurement to optimise programmatic advertising.

The platform scales its AI solution, Koa, by distributing deep learning algorithms, making AI a co-pilot for advertisers.

Mitch Waters, the senior vice president for client services in APAC at The Trade Desk, tells Campaign that Kokai addresses the complexity and real-time decision-making challenges in programmatic media buying by leveraging machine learning and AI's analytical capabilities.

According to Waters, the platform processes complex data sets, providing insights and understanding customer behaviour at a granular level, allowing for targeted campaigns and personalised messages that boost engagement and conversion rates.

Regarding measurement, Kokai introduces innovations in the context of CTV and retail media marketing. It consolidates and simplifies measurement, enabling brands to connect digital ads across the Internet to online and in-store sales across multiple retailers in a single report.

“Retail media has emerged as a significant and new innovation in programmatic, giving advertisers new opportunities and complexities that come with advertising across multiple retail media networks,” explains Waters.

“Due to the fragmented retail advertising landscape, measuring outcomes across retailers can be challenging. The many channels, platforms, and targeting strategies marketers employ contribute to a lack of unified metrics and a comprehensive view of results.”

The skills perspective

For brands and agencies looking to incorporate AI into programmatic advertising, understanding their target audience's needs and maximising AI's effectiveness strategically is critical.

Continuous investment in AI talent, upskilling and reskilling existing employees, and regularly reviewing AI tools to evaluate performance and refine them are crucial to maximise AI's benefits in marketing automation and deliver value to marketers and consumers.

Matt Tindale, the head of LinkedIn Enterprise Marketing Solutions in APAC, tells Campaign that generative AI is rapidly transforming the world of work, sparking interest and conversations on AI.

Tindale notes that marketers are keen to use AI to free up time for more strategic thinking, focusing on strategy and creativity.

For example, discussions on LinkedIn about AI have increased nearly seven times from April 2022 to April 2023, with massive spikes in interest coinciding with the launch of ChatGPT in November, 2022 and the more recent product integrations in existing software packages. Interest has recently spiked even more among marketers, whose engagement around responsible AI has grown 15 times over the same period.

“By removing menial tasks, AI empowers individuals to shift their focus towards critical thinking, problem-solving, and innovation, thus serving as a valuable tool to augment human creativity and potential. For marketers, this might mean removing basic writing tasks or group email content and allowing them to focus on big-picture thinking,” explains Tindale.

“That being said, the more significant focus lies in the pace of change of skills more broadly, which underscores the need for companies to focus on upskilling and reskilling their workers across all areas and for professionals to have a growth mindset. This will help professionals stand out and, more importantly, prepare them for future jobs.”

The sustainability angle

While specific case studies on generative AI's impact on data waste and carbon footprint in programmatic advertising are still limited, the potential lies in creating more personalised ads that reduce the number of ads shown to uninterested audiences.

Reducing the number of ads shown to uninterested audiences will cut wasted resources like energy, water, and materials used in ad production and distribution.

Tom Jones-Barlow, the general manager for APAC at SeenThis, notes that AI's development is still in its early stages across various domains, including the media industry. Extensive experimentation is ongoing to explore AI's potential value in advertising for performance and sustainability.

“The application we are already seeing is in creative production like making it quick, easy and cheap to practically instantly create various image, text and video assets for creative,” Jones-Barlow tells Campaign.

“Traditionally, campaign managers have never had enough creativity to test what works best thoroughly, and now they can. This could improve effectiveness because fewer impressions must be bought to get the same results, minimising CO2 footprints.”

Looking ahead, Jones-Barlow envisions the future of sustainable programmatic advertising to involve further reducing carbon emissions by adopting adaptive streaming solutions.

“We can increasingly expect algorithms and technology to be directed, reduce data consumption, and shorten supply chains in digital marketing,” explains Jones-Barlow.