Perhaps no ethos embodies Silicon Valley more than the idea of “open source”, and its natural converse that argues “walled gardens” ultimately fail because they can’t match all the benefits that come from being “open” and interoperable.
And perhaps no open source digital platform has helped as many businesses – through the power of Search Engine Optimization (SEO) – organically grow as Google Search has over the last 25 years.
So it makes perfect sense that AI optimization (AIO) is the next growth opportunity for every business big and small. Indeed, it feels as though for many businesses the race is on to create and push as much of their information into the frontier AI platforms in hopes that links to their brand appear in billions of AI conversations. And if this works, these businesses believe, they’ll reap the rewards from this new open source customer acquisition channel called AI.
I’m not so sure history will repeat itself here.
Google Search and the Open Web
Using content publishers and brand marketers as examples, the overly simplified evolution of Google Search goes something like this:
In the early years publishers and brands could invest resources to reverse engineer how Google Search worked, learn how queries could be relevant to their business, and, as a result, generate organic traffic, audience, and potential customers. This is the general concept behind SEO.
During this same period, Google invested billions to optimize its role as the “middleman” between publishers and brands, and their target audience and customers. This is the overly simplified description of how search engine marketing (SEM) evolved.
Google’s financial aspirations pushed the company to make it harder for businesses to consistently succeed at the SEO game. Most notably the moving target for SEO success has been Google’s persistent schedule of “search update releases” designed to curb the volume of very low cost audience and customer acquisition traffic publishers and brands could garner. Concurrent with the narrowing of the SEO opportunity, more of the consumer search results experience has become filled with SEM-driven paid marketing placements.
Despite this not-so-subtle shift from SEO to SEM, you can still argue that Google Search remained an open, non-walled-garden environment. Why is this? Well, think of Google Search as a 2-dimensional (2-D) platform.
Search is 2-D because it’s presented in a single user interface, or doorway, that includes a single screen measured by your device’s height and width, and when the user clicks on a search result they are routed to a third dimension that exists beyond Google Search – whether the user gets there via an SEO or SEM link.
So based on the way history has played out in search, you can’t fault publishers and brands from looking at AI platforms as the next big traffic generating “open source” opportunity. But they shouldn’t.
AI Conversations and the New Walled Gardens
At first glance the parallel looks enticing. If you’re a publisher or brand, all your amazing content and intellectual property should be accessible by ChatGPT, Claude, or Gemini, and when it’s relevant to serve up your content in an AI conversation – boom! – users will find it, and happily click links to find more at your website. Not this time.
AI platforms like OpenAI, Anthropic, and Google are actually 3-dimensional, “3-D” versus 2-D like Google Search. What does 3-D mean here? Think of your prompt like a search query, and you enter it into a screen with height and width (2-D), but instead of getting a result that clicks you to a 3rd party for the third dimension, you enter the third dimension right there in the AI’s response.
And the loop continues inside the AI platforms, with each subsequent prompt, response, and ensuing conversation. Heck, you may never leave, get everything you need and never visit any of the sources ChatGPT, Claude, or Gemini used to help answer your questions. Welcome back to the AI walled garden.
What’s a Business to Do?
If you are a publisher, content creator, or IP generator that connects potential customers to a digital service you hope to get paid for, expect AI “walled gardens” to find this information and republish it in conversations - maybe with attribution and a link to your business - but likely not. Until we see how the copyright lawsuits that publishers like the New York Times have brought to the likes of OpenAI and Anthropic, it’s safe to assume your IP will stay in the AI ecosystem where the vast proportion of value will accrue to the AI platforms.
If you are a brand selling a product or service that can’t be fulfilled within the AI conversation itself, you’ll see the gap between AI “optimization” (AIO) and AI “marketing” (AIM) collapse way faster than the gap did between SEO and SEM. Google’s aggressive effort to inject their AI platform Gemini into the “search UI” is a clear sign they are replacing the 2-D model with a more lucrative 3-D model where Gemini subscriptions cover experiences you don’t need to leave Gemini to consume, and brands will continue to pay Google for customer acquisition through AIM instead of SEM. At the same time, OpenAI’s public admission that “ads” will be coming to ChatGPT clearly implies they’ll be working on the same calculus, trying to optimize the subscription revenue users pay them and the AIM revenue 3rd party businesses pay them.
Ultimately for CEOs and business leaders, the important question is determining how quickly AI platforms can disintermediate what you do, particularly if the IP that makes your business valuable becomes fully embedded within these AI environments. For many businesses, as counterintuitive as it sounds, the answer may be going back to building and maintaining their own “walled gardens” in an effort to directly acquire and engage your target customers. And even if that’s the answer for your business, be prepared to migrate a major portion of your marketing spend to AIM in order to grow and maintain your own healthy walled garden.
NextPlay>Forward AI Disclaimer: I very actively use artificial intelligence and large language models to generate the content you read here, but I do review it and edit it to make sure it can be generally useful to people who read it. Keep in mind that AI can make mistakes - check important information. Let me know if I make any errors and I will correct them.