To grow a business, you need customers to pay you money for a good or a service. And historically, these customers have been people. You know, actual humans with wallets and opinions. Even if the end-user is a business (B2B), it’s ultimately the people at the business that make the buying decision.
But times are changing. Not only do companies have to continuously innovate on customer acquisition strategies for humans, but soon, they will also have to create a playbook for marketing to AI agents making decisions on behalf of customers.
Let’s survey the scene.
On Monday, the WSJ published an article about how Meta plans to leverage AI to automate content creation on its advertising. According to Meta in the WSJ article, all you need to do is, “present an image of the product it wants to promote along with a budgetary goal, and AI would create the entire ad, including imagery, video and text.”
Wow.
It’s not a far-fetched statement to say that creating advertisements is a time-consuming process that takes a specialized skill set. But here comes Meta, saying that if you provide a picture of what you want to advertise, they’ll create multiple formats of an advertisement in an instant. Assuming the AI content is high-quality, why wouldn’t you advertise on Meta? You saved a ton of time (and not to mention, money) by having someone else (Meta) create the content for you.
What makes this even more intriguing is that back in 2022, Meta introduced Advantage+ Shopping Campaigns, a new type of advertising campaign that simplifies the media buying process by letting their AI do the work. You set up a few basic parameters, upload image and video content, and then Meta figures out which content should be shown to potential customers.
They turned digital advertising into a BYOC (bring your own content) situation. But soon enough (Meta says by the end of the year), you won’t need to bring your own content as Meta will have you covered! For what it’s worth, Google has also introduced similar hands-off-the-keyboard advertising campaign structures (Performance Max) to let their AI do the work instead of a human operator. Although Google advertising operates differently than Meta (paid search versus paid social), I imagine that Google is working on a way to further automate campaign execution beyond what they currently have in place.
Furthermore, the shift to AI becoming the media buyer lines up well with what we are seeing on the customer side.
In some cases, I anticipate that the definition of a “customer” will change from a human to an AI agent. As we’ve discussed in AI + Shopping, we may see our own personal AI shopper make routine purchases like grocery and household necessities on our behalf, based on pre-defined guardrails. If an algorithm becomes the shopper, it becomes even clearer to me that marketers and operators should rely on AI for digital advertising content and execution. I trust an advertising algorithm to stay up to speed on the nuances of marketing to AI shoppers more than a human on a keyboard (no offense, direct response marketers).
And now that Meta is claiming they will automate the content generation side, the end-to-end digital advertising process is taken care of. Let the algorithms go out and acquire the other algorithms, who are essentially the new customers.
What are the consequences of this? My guess is that advertising content quality declines. Because AI advertising algorithms are looking to appeal to AI shopping agents, the images, videos, and copy may start to be designed for an AI shopper to pick up, not a human. I think it’s still too early to have a strong sense of how this will play out if we let AI replace our shopping and our direct response advertising, but my guess is that advertisements will be choppier and less cohesive. That’s because they’ll be focused on fitting in plenty of content to attract an AI shopper, over having a cohesive storytelling element that appeals to humans.
For better or for worse, humans may not be the primary customer of digital advertising soon.
…
But AI won’t change everything. AI can’t automate a fun, in-person event!
Enter New York Tech Week. There are over 1,000 scheduled events this week across Manhattan and Brooklyn, spanning every sub-industry of Tech.
It’s the classic event marketing playbook. Bring like-minded individuals together (ideally in-person) and host a fun event that fosters community and connection. For Tech Week, the events are mostly hosted by Tech companies (Amazon), investors (a16z), law firms (Fenwick), and niche networking groups (Andrew’s Mixers). Events range from happy hours to dinners and even padel tournaments.
On Wednesday, I stopped by Perplexity’s Curiosity Café, a fancy name for what was a food truck handing out free coffee. As I waited in the long line (I guess a lot of people also like AI and free coffee), I struck up a conversation with a Perplexity product manager. We chatted on a few of their new products, ranging from Shopping to flight booking. Over the past few days, I’ve been switching up my prompts from ChatGPT to Perplexity. I like what I’m seeing, and I may sign up for a Perplexity Pro subscription!
Side note: try typing the same exact prompt into a few different LLMs (ChatGPT, Perplexity, Claude) and compare outputs. It’s fascinating to observe the differences in output regarding content, style, and tone.
So, what’s the moral of the story? Even the people building technology that will automate commerce and media choose to spend at least some of their marketing dollars on meeting up in person with potential customers.