Kevin Goodwin, VP of Performance Marketing
Jen Thomas, Creative Director
Anna Otieno, Head of Research & Insights
Michael Goodwin, Chief of Staff
Over the last year, the public discourse has been flooded with conversations about generative AI - what it is, how it should be used, and to what extent it will (or won’t) change us as a society. Because the implications are widespread, stretching from the tech sector to the education system, we believe it’s important to address some potential impacts on marketers and how we, as an agency, are thinking about generative AI.
Last summer, a handful of AI image generators were made publicly available, giving the masses a taste of what generative AI tools can do. The technology quickly entered the American lexicon, with shows like Last Week Tonight with John Oliver shining a spotlight on Midjourney and DALL-E. Then, in November 2022, OpenAI unveiled ChatGPT: a pre-trained, generative chat tool that uses Natural Language Processing (NLP), drawing from textbooks, websites, and articles available on the internet in order to model its own language in its responses. ChatGPT can give natural answers to questions in a conversational tone, and can generate stories, poems, debug code, recommend chords and lyrics, and more. In early February, New York Times tech columnist Kevin Roose reported that ChatGPT had 30 million users and 5 million daily visits, just two months post-launch.
Since its release, ChatGPT has escalated the national conversation around artificial intelligence; a February survey published by YouGov found that 46% of Americans have heard about ChatGPT, but their feelings about the technology are mixed at best. 46% of respondents believe that AI will negatively impact jobs (17% disagree), and 36% say it will be “bad for society“ (13% disagree).
And as far as businesses and institutions go? They’re starting to tackle big questions about AI best practices, protocols, and use cases. Bloomberg recently introduced a financial tool called BloombergGPT, Universities are adding Generative AI FAQs to their websites, and the White House published a Blueprint for an AI Bill of Rights, detailing principles that “should guide the design, use, and deployment of automated systems to protect the American public in the age of artificial intelligence.”
The list goes on, but let’s turn now to marketing. In this piece, we’re going to explore ways in which generative AI stands to impact marketers, use cases for the technology offered by New Engen leaders, and our agency mindset as we enter this next technological frontier.
Publishers have been building AI into their ad platforms for some time now. In late 2021, Google introduced Performance Max - an automated tool that serves ads across Google’s ad inventory from within a single campaign. Shortly after, in August 2022, Meta unveiled its Advantage+ shopping campaigns, which use AI to streamline the process of ad creation and optimization. Basically, “Our highly automated marketing future has arrived,” says Kevin Goodwin, VP of Performance Marketing at New Engen.
And now, just a few months later, these same publishers are racing to incorporate generative AI tools into their product offerings. Microsoft launched “the new Bing,” an AI-powered search engine that the company believes will improve audience targeting and investment returns for brands. Last month, New Engen was invited to attend a Microsoft webinar called “Introducing the New Bing,” which highlighted how AI can enhance and complement traditional search (but not supplant it, according to the webinar’s host). Microsoft says that, for now, no changes will be made to the ad auction and ads for the new Bing will be automatically created within existing campaigns. The most significant difference, they claim, will be in improved relevancy, with more qualified consumers discovering a brand’s content and ads.
In light of Bing’s AI evolution, Google has been hard-pressed to advance its own search engine and, in March, released a chatbot the company is calling Bard. The technology is still in beta, and time will tell what impacts it might have on the platform's advertisers.
Zooming Out: Marketers Can Relax About Chatbots & the SERP
"A lot of people are worried that these chatbots will lead to singular query results that will reduce the opportunity for ads, completely alter the way the search engine results page (SERP) appears, or both. But it should give marketers peace of mind to remember that Google and Meta are incentivized to create great ad products that generate revenue for their shareholders. Search was 56% of Google's revenue in Q4 2022 - it benefits them to use AI in a way that allows advertisers to spend more.
What we can expect to see is greater curation, enabled by chatbots in the SERP. An early tactical indication of the shift in search is the evolution of match types. Google has already done away with broad match modified and is pushing broad match harder than ever. As we move to more voice-based and long-form queries, the ability to build an effective and scalable keyword strategy with tools like exact match becomes near impossible. I would expect match types, or at least any but broad, to be almost completely obsolete within the next 3-5 years if not sooner."
- Kevin Goodwin, VP of Performance Marketing
Meta has, of course, also thrown its hat into the ring. In a recent interview with Nikkei Asia, Meta’s CTO Andrew Bosworth remarked that the company was developing a generative AI solution for ad creation that it plans to launch by the end of this year. Bosworth explained that the technology would be able to generate ad creative tailored to different audiences across Facebook and Instagram, simplifying the advertising process and making it more widely accessible.
New Engen leaders anticipate that the first iteration of Meta’s generative AI tool will level up its existing dynamic creative and automated ad formats, like Slideshows. Kevin Goodwin shares his predictions:
Prediction 1: Meta Will Focus on Feeds
“Meta is likely to focus on feeds in the near term, given the number of ad dollars invested in feed-based products. What this might look like is a tool in which advertisers provide their brand colors, input images, and a set of rules, and in turn, Meta will produce automated feed manipulations like backgrounds and overlays using AI.”
Prediction 2: Recommended Videos & Images Aren’t Far Off
“I believe very quickly we'll see Meta attempt to create 'recommended ads' that splice together top-performing products and assets already within your account. Imagine you have an ad featuring a shoe on a textured background that is performing well. You can click a button that says 'Generate Optimized Variation,' and Meta manipulates your image - maybe changes the background texture, maybe includes some text on the image, maybe puts AI-generated 'people' in your ad, etc."
The internet is having a field day with AI-generated content. AI images of figures like the Pope and Donald Trump have stirred the pot in recent weeks, and college professors are trolling their colleagues with academic papers produced by chatbots. But politics and pop culture aside, some major retailers have made recent headlines for their use of AI-produced marketing collateral.
On March 20th, Coca-Cola launched the “Create Real Magic” marketing campaign to promote its new AI platform, developed by OpenAI and Bain & Company. In a press release, artists were encouraged to visit the platform’s site and generate AI artwork using “iconic creative assets from the Coca-Cola archives” for the chance to be featured on Coke’s digital billboards in New York and London. Two days later, Levi Strauss & Co. announced it would be partnering with Lalaland.ai to build “customized AI-generated models.” As in fashion models, not language models.
Retailers are also jumping on AI to differentiate their gifting experiences, says Anna Otieno, New Engen’s Head of Research. She explains, “Brands are increasingly incorporating generative AI to optimize the customer experience – from search to delivery. Being understood is a top value for consumer interest and engagement. AI-powered gifting generators, for example, are popping up more but are in the development stages of personalization. A tool that helps customers find the perfect gift – and links directly to hyper-personalized products – is already steps ahead of current options. It’s like a personal shopper that goes beyond the usual ‘recommended for you’ features.”
AI and machine learning continue to stir up the industry, and several applications have the potential to deliver serious, positive impacts at scale: AI-powered ad optimization, smart tagging, templatization, video cutting, image generation and editing, automatic cropping, and text generation, to name a few. What’s more, creative professionals are already using generative AI tools like ChatGPT, DALL-E, and RunwayML in their work. But all that being said, the potential benefits of these tools are counterbalanced by reasonable concerns surrounding content quality, security, and liability.
Test & Learn
For the past several years, New Engen Creative Director Jen Thomas has been working with internal analytics and media teams to assess possible use cases for AI. She tells us,
I’ve approached AI like I would any other tool in my line of work - by embracing it, but also attempting to fairly evaluate its strengths, weaknesses, ethical considerations, and potential near-term evolutions.
Jen and other New Engen creatives have found that quality concerns surrounding AI tend to limit the success of content creation. “In research and discovery with images and editing, AI hasn’t led to significantly better speed, volume, or quality of output than a person working manually, at least so far,” Jen explains. She adds that, over time, an undesirable “stock” quality can become apparent in AI-generated content. Ultimately, the level of tailoring required to achieve specific outputs is still inaccessible to many, and the QA and editing processes continue to require a trained eye.
ChatGPT and Copywriting
With thoughtful prompting, ChatGPT has shown intriguing strengths, like a facility for loose, engaging, brand-focused language. However, it has also revealed weaknesses, like basic inaccuracies and stiffness in the kind of product-specific, feature-focused language that digital advertising often requires. Certain industries like nutrition and wellness may encounter additional challenges and require an even greater need for human handling - due to the natural language sound of the responses, QA testers must know products at an individual level to be certain any health claims are accurate enough to publish.
This demand for skilled prompters has created a job market for Prompt Engineers, and prompt-sharing is becoming something of a sport for people seeking the right language to describe the output they want.
What to Look Out For
New Engen leaders believe that, right now, the most compelling use case for AI is creative analysis and insights. Machine learning options, like smart tagging and image classification, are where Jen and others see the most potential in a paid media environment. And regarding ChatGPT specifically? Anna Otieno puts it this way,
At an estimated 100 million monthly users, it’s no question that ChatGPT will impact (and has already impacted) how people search, learn, interact, create, work and access information. As an agency, it’s an active learning process, where brands will test and discover organizational uses and consumer-facing tools. The rise of ChatGPT will also call for research and policy around privacy, bias and ethical use. The sooner brands get ahead of that, the safer the adoption for all parties.
The topic of generative AI is so massive, has so many angles and applications, that we can only attempt to scratch the surface with this piece. At New Engen, our approach to AI is the same as it has been with any new technology - welcome it, test into it, and find the best use cases for our clients. And we'll go step further by applying this framework equally to third-party technologies and those we are incubating in-house (but more on that later). Right now, we see the greatest short-term potential for AI to expand our creative analysis and insights, and help evolve media productivity and output.
Let’s end with some perspective from Kevin Goodwin, who offered his thoughts in a recent AiThority article,
Hyper-automation is an opportunity to refocus our industry toward producing truly inspiring creative, and the technology that enables it. The machines, platforms, and algorithms can handle optimization and measurement—and marketers can return to a time when advertising and media conversation was about creative that truly moves people.
Keep an eye out for more AI news and stay in the know by subscribing to our newsletter and following us on LinkedIn.