What’s Really Possible; The New Home Takeover

What’s Really Possible; The New Home Takeover

The POSSIBLE conference in Miami is abuzz with discussions about agentic AI in programmatic advertising. Marketers and industry experts are eager to separate the hype from reality. While agentic AI tools exist, their current capabilities and limitations need to be examined.

The Current State of Agentic AI in Programmatic Advertising

Agentic AI tools are being used in programmatic advertising, but their applications are currently limited. These tools mainly help agencies and sellers automate workflows or formerly manual campaign setups. According to Lara Koenig, VP of global strategy and partnerships at MiQ, “We’re still a ways from truly autonomous agentic buying.” This assessment is shared by many attendees at the POSSIBLE conference.

Automating Workflows, Not Buying

The current state of agentic AI in programmatic advertising is focused on automating workflows and campaign setups. This includes tasks such as data analysis, ad placement, and bid optimization. While these tasks are important, they are not the same as truly autonomous agentic buying. The industry is still in the early stages of developing agentic AI tools that can make decisions on their own.

The Hype vs. Reality

There is a significant gap between the hype surrounding agentic AI and its current capabilities. Many marketers and industry experts are skeptical about the near-term potential of truly autonomous agentic buying. The reality is that agentic AI tools are not yet capable of making complex decisions on their own. They require significant human oversight and intervention.

A Nuanced Understanding

It’s essential to have a nuanced understanding of the capabilities and limitations of agentic AI tools. While these tools have the potential to revolutionize programmatic advertising, they are not a silver bullet. Advertisers and agencies need to understand what these tools can and cannot do.

Early Examples of Agent-to-Agent Deals

There are early examples of agent-to-agent deals, such as AI startups like Newton Research working with ad agencies like RPA. These deals are focused on direct-sold campaign automation rather than open-auction programmatic. However, they demonstrate the potential for agentic AI tools to automate complex tasks.

Direct-Sold Campaign Automation

Direct-sold campaign automation is an area where agentic AI tools are being used. These tools can automate tasks such as ad placement, targeting, and bidding. This can help advertisers and agencies optimize their campaigns and improve efficiency.

Building Toward an Agent-to-Agent Future

Companies like Newton Research are building toward an agent-to-agent future for programmatic. According to John Hoctor, CEO of Newton Research, “We’re training our models on the wealth of institutional knowledge compiled by the IAB and other advertising trade groups – including from thought leadership, case studies, and bits of programmatic-specific code.” This approach has the potential to enable truly autonomous agentic buying.

The Role of Institutional Knowledge

Institutional knowledge compiled by the IAB and other advertising trade groups plays a critical role in training AI models. This knowledge includes thought leadership reports, content marketing case studies, and bits of programmatic-specific code. By leveraging this knowledge, AI models can learn from the experiences of others and improve over time.

The Unconventional Data Sources Powering Agentic AI

Agentic AI tools are being trained on unconventional data sources, such as thought leadership reports, content marketing case studies, and bits of programmatic-specific code. This approach has the potential to enable agentic AI tools to learn from a wide range of sources and improve over time.

A New Approach to Data

The use of unconventional data sources to train AI models is a new approach to data. This approach recognizes that traditional data sources may not be sufficient to enable agentic AI tools. By leveraging a wide range of data sources, AI models can learn from multiple perspectives and improve over time.

The Black Box Problem

One of the potential risks and challenges associated with agentic AI is the lack of transparency and accountability in AI-driven decision-making. This is often referred to as the “black box problem.” Advertisers and agencies need to understand how agentic AI tools make decisions and be able to audit their performance.

The Need for Transparency

Transparency is essential when it comes to agentic AI tools. Advertisers and agencies need to understand how these tools work and be able to audit their performance. This can help build trust and ensure that agentic AI tools are being used effectively.

The Talent Gap

There is a need for advertisers and agencies to upskill and reskill their workforce to effectively work with agentic AI tools and leverage their potential. This includes developing new skills, such as data analysis and AI programming.

Upskilling and Reskilling

Upskilling and reskilling are essential for advertisers and agencies that want to leverage the potential of agentic AI tools. This includes developing new skills, such as data analysis and AI programming. It also requires a deep understanding of how agentic AI tools work and how to use them effectively.

Conclusion

In conclusion, agentic AI tools have the potential to revolutionize programmatic advertising. However, their current capabilities and limitations need to be examined. Advertisers and agencies need to understand what these tools can and cannot do. By leveraging unconventional data sources and developing new skills, advertisers and agencies can unlock the full potential of agentic AI tools.

For more information on the intersection of advertising and technology, check out A New Standard For Transparency In CTV: What Advertisers Should Expect. Additionally, Supplement Brand Ritual Taps Chord To Help Understand Its Historical Data provides insights into how brands can leverage data to improve their marketing efforts.