AI-driven search tools like Perplexity and Deep Research are reshaping online shopping and SEO, challenging traditional product review sites and pushing publishers to optimize for generative AI.
These past few weeks have seen a whirlwind of “deep” announcements: DeepSeek, Deep Research, and DeepSeek’s R1 integration with Perplexity. Much of the attention has focused on the impact of these tools on research papers, but they could also transform how consumers make high—consideration purchases. And that, in turn, can have a big impact on the SEO industry and certain types of publishers.
Let’s start with a quick look at some of the AI-search comparisons.
We’re all familiar with Google’s Search Generative Experience (SGE), which integrates AI-generated summaries in response to user queries. Importantly, for Google, AI Summaries keep a core advertising model in place—sponsored Links appear above or below the AI summary, and they prioritize Product carousels and Shopping ads.
Perplexity takes a more comprehensive approach. When I asked for recommendations for running shoes for people with bad knees that cost less than $100 and come in blue, it explained shoe features that are appropriate for bad needs and generated a detailed comparison chart that included citations to authoritative sources (e.g., Runner’s World and Run to the Finish) as well as Reddit reviews. The results finished with direct links to retailers’ product pages that offered the best prices for each shoe model. Perplexity says it can find hidden deals by scanning prices across Amazon, Best Buy, Shopify merchants, and other retailers to find the best available deals.
OpenAI’s Deep Research is also useful to consumers. It’s only available to Pro users, but the company promised to roll the feature out more fully. In the YouTube video announcing Deep Research, the presenter wants to buy skis for a last-minute Japan trip. He asks for recommendations based on his preferences: advanced skier, all-mountain with powder capability, long skis for his height, and a specific color palette. Deep Research scours the internet, just as a thorough consumer would do when researching a major purchase, and surfaces a suggestion that matches the ones he has at home.
So, while Google SGE integrates AI into its existing search-and-display model, Perplexity and Deep Research offer end-to-end shopping assistance, from initial research through price comparison to one-click buying.
What’s Lost in the End-to-End Model?
The Perplexity and Deep Research approach raises questions about the future of specialized product review sites. I’ll give you an example: When planning a kitchen remodel for my 200-year-old Vermont farmhouse, choosing the proper range was critical. Repairs in rural areas can be challenging and expensive, so I wanted a brand and model that was highly reliable (I read the reviews on every retailer’s site that offers them). Still, I wanted more assurance, so I was more than happy to pay for a subscription to Consumer Reports. Out of curiosity, I asked Perplexity to recommend ranges based on these parameters— reliability, ease of repair in rural Vermont, and positive consumer reviews—and it recommended the same models as Consumer Reports.
Suppose these search tools can efficiently synthesize product information, expert reviews, and user experiences and package them into digestible charts. What does that mean for the future of specialized product review sites?
Beyond the recommendations themselves, something else may be lost: serendipitous discovery and impulse buys. As I researched ranges on Home Depot, ads for dishwashers with cool racks for stacking dishes caught my eye, leading to an unplanned but valuable addition to my renovation. Will we miss these unexpected discoveries in an AI-curated world that delivers precisely what we ask for?
An Inevitable Switch
Let’s say I love my serendipitous finds as much as retailers enjoy the opportunity to upsell my purchases. Can I hold onto them?
Not a chance. Google has already integrated AI summaries into 47% of searches, and that percentage is likely to grow. As consumers discover the advantages of doing research on tools like Perplexity and Deep Research, will Google get into the game and offer end-to-end experiences as well?
Reading the tea leaves, perhaps competition in search is one reason Google’s CEO warned employees that AI will make 2025 a challenging year for the company. And maybe it’s why the company just announced it will invest another $75 billion in AI this year.
This potential shift should have an impact all publishers and businesses that rely on search traffic, forcing them to add generative AI engine optimization (GEO) to their strategies.
SEO v. GEO
GEO is what it sounds like: an attempt to create digital copy that AI search engines will use when serving up results. However, SEO and GEO require fundamentally different optimization strategies, which Search Engine Land effectively lays out. For instance, while SEO optimizes traditional search engines that provide users with links in response to their queries, GEO optimizes content for AI systems to synthesize and prioritize information, generating quick and comprehensive responses.
While SEO optimizes meta tags and content with keywords to improve search rankings, GEO is all about ensuring that content is clear and contextually relevant so that AI algorithms can use it to generate accurate and comprehensive responses.
Citations of authoritative sources (e.g., Runner’s World in running shoe queries) now carry increased weight—three-quarters of websites cited in AI overviews rank in Google’s top 12 organic results. This suggests that high-ranking, authoritative content gets preferential treatment in AI summaries.
The Future of Search
The transition from SEO to GEO isn’t optional, it’s inevitable. As Google integrates AI summaries into more searches and tools like Perplexity, SearchGPT (announced last week), and Deep Research gain traction, publishers and marketers will need to adapt their content strategies. Traditional SEO will continue to be important, but success in gaining referral traffic will increasingly depend on content that AI systems can synthesize.
While this shift will upend long-standing ways of finding information and products online, consumers will get faster, more complete search results. For publishers and marketers, adapting to GEO will be table stakes for visibility in AI-driven search.