News | Generative AI in Advertising: A Road Map for Scalable Implementation
With 2024 on the horizon, advertisers are evaluating how to invest in AI – particularly, generative AI. Reliable and practical applications, however, can be hard to find.
How should a business begin thinking about its generative AI strategy against this buzzy backdrop?
Start with your business objectives and how AI can amplify your unique value proposition as an organization. Most of the advertisers I work with seek, above all, to keep their promises to their clients. When adopting a new technology, they apply a baseline of sound fundamentals. These same core principles can be applied toward your AI approach.
Having started my digital advertising career almost 20 years ago, I come from the first wave of automation in digital marketing.
I’ve learned many lessons from pioneering newer technologies as president and co-founder of Fluency. Our company uses robotic process automation (RPA) to help advertisers deliver an unlimited number of locally targeted campaigns at scale.
One of the biggest lessons I’ve learned is how to get people to use new tech. It has to save users time or boost performance (preferably both). You could simplify these value-adds further by capturing them in a single word: productivity. If new tech increases productivity, it will succeed.
As an industry, our grip on how generative AI increases productivity in the here and now is slippery. This article will lay out a runway for the most promising practical applications of generative AI for advertisers, particularly those who seek performance and efficiency gains at scale.
State of the union
If we have learned one thing from the last year of AI development, it’s that AI success depends on human oversight.
For example, content writing was one of the most promising areas of generative AI’s application to digital marketing. ChatGPT promised a world where copywriting might be a simple matter of a few elegant prompts to an AI assistant. But even with careful prompts, unassisted writing has proven risky. Aside from more general problems like hallucinations, there are compliance risks with truth-in-advertising laws and other legal frameworks, such as the Fair Housing Act. Moreover, concerns about generative AI reflecting societal bias have not yet been satisfactorily resolved.
The capacity to work with and interpret data sets was also a much-touted feature of generative AI. However, generative AI’s output is fundamentally nondeterministic. As such, its capability to support and enhance data interpretation must be overseen by experts and harnessed within well-defined use cases.
Our industry is creative and driven by strategic client-agency relationships built on trust, which adds to the inherent risks. Understandably, clients want outcomes they can count on. They also want accountability for how those outcomes are arrived at. With generative AI’s nondeterministic outputs, leaving this trail of breadcrumbs is impossible.
The obvious solution is to have safeguards – to carefully comb through the output to verify accuracy and compliance. But this process can be resource intensive, undermining the efficiency that AI is meant to bring.
Practical solutions for advertisers
Despite these limitations and key considerations, we are observing some effective current and emerging use cases for generative AI. And we expect these will continue to grow in the years ahead. They fall into three categories: generative AI as interactive help, content inspiration and value articulation.
First, we’ve seen generative AI work as a highly advanced chatbot. Trained on a corpus of a company’s data, it can provide instant and accurate assistance to clients – even on complicated questions. It can also serve as an internal knowledge-management tool. For content inspiration, generative AI is an effective sandbox for roughing out drafts, ideating and proposing possible directions.
Value articulation may be the most overlooked of generative AI’s current quick wins. Communicating value across large portfolios of clients will take a significant amount of an account strategist’s time. It requires more-or-less constant communication and reminders of why a business is better than its competitors. But generative AI – particularly, its video and image capabilities – offer a huge lift, building immersive reports that have audio-visual components, walking prospects as well as established clients through the value that your business can bring to them. Account strategists are typically giving attention to around 30 accounts each. Here, the value of automation comes into play. Automating value articulation through GAI liberates a great deal of their time, allowing them to take on more and more accounts and bring in more revenue.
We’ve found generative AI very capable of producing reports that can articulate a company’s value proposition, describe its market, convey movements in the global economy and package all this information into a video narrated by an AI avatar that can be sent to its clients.
What to expect
These strategies are capable of adding value to your business immediately. They’re safe areas of experimentation that don’t engage the aforementioned risks of nondeterministic output, bias and noncompliance. In the coming year, we will likely see more pragmatic use cases emerging from these strategies. Building up from proven and secure approaches is perhaps the surest way to reach the full potential of generative AI – iterating on them will build the necessary literacy and facility with generative AI to begin branching out into other areas of value creation.
We will also begin to see the emergence of more legal regulatory frameworks to address the aforementioned risks. As might be expected with a new technology, there’s a current paucity of regulation. And while this might seem favorable to business, in reality, it makes using generative AI murkier than it ought to be. We might see regulation tame some of the current concerns around unassisted content creation, as well as the emergence of more special ad categories.
A useful tool among others
As we look to 2024, generative AI should be viewed primarily as a tool, not a magic bullet. Like any tool, its application should be dictated by the needs of your business and your core value proposition. From that lens, the central question generative AI will pose isn’t “How do I use this?,” but “How can I use this to add more value to my customers?”