AI agents are engaging in searching for brands at a scale that’s equal to a third of equivalent organic search, according to a report from digital marketing company BrightEdge published Wednesday.
BrightEdge defines an AI agent as any AI tool that does work on your behalf. For example, if you ask ChatGPT to find the most cost-effective bikes for teenagers to ride on gravel, OpenAI will use its live web browsing feature to look across bicycle brands to find the information. ChatGPT isn’t the only AI bot that searches across websites to find necessary info — other popular AI search agents include Google’s Gemini, Perplexity, Microsoft Copilot and Claude.
Truly agentic AI goes beyond these chatbots, too — this report doesn’t include the recent launch of ChatGPT Agent, which can browse the internet in a virtual desktop environment, going beyond research and doing things like booking you flights or ordering you a pizza.
«First, we’ll be delegating more decision-making and prioritization to LLMs — similar to how Google became the main entry point to the web, but this time with multiple players like Google, ChatGPT and potentially others,» BrightEdge CEO Jim Yu said in a statement. «Second, we’ll gain the ability to match and research products and services at a level of detail that simply wasn’t practical before. The flip side is that what you see will depend on what the AI finds and prioritizes, which will have implications for transparency, diversity of results and accuracy.»
The rise of AI means humans are offloading more informational synthesis to machines. This makes research and other types of work much faster and easier (though you should always double check anything an AI tool suggests).
It also means that you can now find products for your very specific requests, rather than trying to figure it out for yourself after reading reviews and internet comments.
Optimizing for AI
Marketers and search engine optimization experts are becoming keenly aware of this. Instead of optimizing for Google Search, as they did in the past, there’s an emerging pivot towards AI optimization. For instance, in the past, if a shoe brand wanted to increase sales, it would try to get to the top of Google Search results for keywords like «shoes» or «sneakers.» This was done via search engine optimization, or SEO.
But in the age of AI, brands need to try and find a way to get their products recommended by ChatGPT and other AI chatbots. Doing so requires some SEO tricks, and also generative engine optimization, known as GEO.
Since AIs aim to understand user intent, brands have to optimize their online content so that it can be easily read by AI, which often considers more complex variables and structured data. Brands will still want to make sure potential customers see their products first.
«History shows that gaming the algorithm never leads to high-quality results,» said Yu. «At some point, the company supplying the answers must step in to address it or risk losing users to better alternatives.»
AI chatbots are vulnerable to prompt injection, a malicious method of tricking AI chatbots into altering their responses. It’s something that Yu acknowledges is a risk, although AI companies will likely build more safety nets to mitigate this.
«The difference is that AI systems have the potential to cross-check information and apply richer quality signals than traditional search,» said Yu. «At the core, brands still need to focus on creating trustworthy, authoritative content and monitoring how they’re represented in AI-driven answers.»
(Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
