AI will also appear in the search business. Or so they say. As Google’s reputation continues to deteriorate and tools like ChatGPT, Google Gemini, and Microsoft Copilot seem to continue to improve, we are propelled toward new ways to find and consume information online. It looks like it is. Companies like Perplexity and You.com are touting themselves as the next generation of search products, and even Google and Bing are making big bets that AI is the future of search. Goodbye, 10 blue links. Hello. Direct answers to all my weird questions about the world.
But what you need to understand about search engines is that they have many functions. For everyone who uses Google to find important, hard-to-access scientific information, to find their email inbox, to visit his website at Walmart, to remember the president before Hoover The number of people using Google has increased by an order of magnitude. And the fact that I like the most is that every year a huge number of people go to Google and type “google” into the search box. We mostly talk about Google as a research tool, but in reality, everything imaginable he is asked to do billions of times a day.
So the real question for all these would-be Google killers isn’t how well they can find information. What matters is how well Google does everything it does. So I decided to put some of the best new AI products to the test. According to SEO research firm Ahrefs, we took an up-to-date list of the most Googled queries and questions and connected them to various AI tools. In some cases, we’ve found that these language model-based bots are really more helpful than Google’s results page. But for the most part, I’ve learned exactly how difficult it is to replace Google at the center of the web, whether it’s with AI or something else.
People who work in search always say that there are basically three types of queries. The first and most common is navigation. For navigation, simply enter the name of the website to access her website. Almost all of Google’s top queries are navigation his queries, from “youtube” to “wordle” to “yahoo mail”. In fact, the search engine’s main job is to direct users to her website.
In reality, a search engine’s main job is to direct users to websites.
AI search engines are generally worse than Google when it comes to navigational queries. When you do a Google navigation search, it’s very rare that the first result isn’t what you’re looking for. Admittedly, it’s odd that Google would show all of these results when what they’re actually supposed to be doing is just directing users directly to Amazon. com is fine, but it’s fast and rarely wrong. AI bots, on the other hand, prefer to think for a few seconds and then provide a ton of semi-useful information about your company when all you need is a link. Some items did not link to amazon.com.
I don’t hate the additional information, but I hate the way it’s done. length These AI tools are necessary to get what you want. Waiting 10 seconds for three paragraphs of generated text about Home Depot is not the solution. I just want a link to Home Depot. Google wins that race every time.
The next most common type of search is informational queries. That is, you want to know something specific for which there is a single correct answer. “NFL scores” is a very popular information query. “What time is it?” is another question. The same goes for “weather”. It doesn’t matter who gave you the score, time or temperature, that’s what you need to know.
Here, the results are displayed across the map. For real-time things like sports scores, AI cannot be trusted. Both You.com and Perplexity frequently provided outdated information, while Copilot usually provided correct information. Not only does Google do it right, they usually pop up widgets with other stats and information, but this one is better than most. The same goes for anything that requires a specific location or context. Google probably has information about you, but AI bots have very little.
As for the more evergreen information, like “How many weeks in a year?” and “When is Mother’s Day?” everything I tested was accurate. In many cases, I actually preferred the AI’s answers. It adds a bit of helpful context to the AI’s answers. But I don’t know how often I should trust it. Google said that a year has 52.1429 weeks, but You.com actually explains that there are 52 weeks and 1 day, and in a leap year there is 1 extra day. did. More useful than just 52.1429. But then Perplexity told me that actually in normal years he has 52 weeks and in leap years he has 52 weeks and 1 day. It then directly contradicts itself two sentences later. This is the whole answer. Try to understand it:
A normal year has about 52 weeks, but a leap year has 52 weeks and one day. To be more precise, a typical year for him actually consists of 52.143 weeks. This means there is one extra day in a normal year. Leap years, on the other hand, occur once every four years, with some exceptions, and have 52 weeks and 2 days. This difference in the number of weeks is due to the fact that in ordinary years one day is added, and in his February in leap years one day is added.
After further research, I was convinced that what You.com said was the answer. However, this takes too much time and requires fact-checking the search, which defeats the purpose of providing a useful summary. Google continues to win here with just one thing he does: speed.
However, there is one subgenre of information queries that is quite the opposite. I call this a “buried information query.” The best example I can give is the very popular query “How to take a screenshot on Mac”. There are a million pages of answers on the Internet. Just Cmd-Shift-3 to view the entire screen or Cmd-Shift-4 to capture a selection. You’re welcome – but that information is usually buried below. Lots of ads and SEO crap. All the AI tools I’ve tried, including Google’s own Search Generative Experience, simply take information and provide it directly. This is fantastic!
Are there complex issues inherent in this that threaten the business model and structure of the web? Yes! But as a pure search experience, it’s much better. We got similar results when we asked for information that is easy to know but hard to find, such as ingredient substitutions, coffee ratios, and headphones’ waterproof ratings.
Now let’s talk about the third type of Google search: exploration queries. These questions do not have a single answer, but instead are the beginning of a learning process. The most popular list counts search queries like “how to tie a necktie,” “why chainsaws were invented,” and “what is tiktok.” If you’ve ever Googled the name of a musician you just heard, or looked up things like “Helena Things to Do in Montana” or “History of NASA,” you’ve been exploring. According to the rankings, these are not the main reasons people use Google. However, this is the moment when AI search engines shine.
Wait, why? was Was the chainsaw invented? The first officer gave us a multi-part answer about its medical origins before explaining the evolution of the technology and its eventual adoption by lumberjacks. He also provided 8 very useful links for further reading. Perplexity gave a much shorter answer, but it also included some cool images of old chainsaws and a link to YouTube commentary on the subject. Google’s results contained many of the same links, but without any compositing. Even the generative search yielded only very basic things.
My favorite thing about the AI engine is quoting. Companies like Perplexity and You.com are getting progressively better at linking to the source (often inline). This means that if you come across a particular fact that interests you, you can go directly to the source. It doesn’t always provide enough sources or place them in the right places, but this is a good and helpful trend.
One experience I had while doing these tests was actually the most eye-opening of all. The most searched question on Google is simple: “What to watch?” Google has a special page design for this purpose, with posters featuring “popular articles” such as: Dune: Part 2 and imaginary; “For you” for me dead pool and stop and light; Next, options by popular titles and genres. None of the AI search engines produced similar results. Copilot listed his five popular movies. Perplexity offered some seemingly random options. girls 5 eva to manhunt to general; you.com gave me a ton of outdated information and suggested I watch “14 Best Netflix Original Movies” without telling me what it was about.
AI is the right idea, chatbots are the wrong interface
In this case, AI is the right idea. You don’t need a lot of links. I need answers to my questions. But chatbot is the wrong interface. The same goes for search results pages. Google clearly knows that this is the most frequently asked question on the platform and was able to design something that works better.
In some ways, this is a perfect summary of the current situation. For at least some web searches, generative AI may be a better tool than decades-old search technology. But modern search engines are more than just pages of links. They are more like miniature operating systems. You can answer questions directly, and it has built-in calculators, converters, flight pickers, and all sorts of other tools to help you get to your destination in just a click or two. According to these graphs, the purpose of most search queries is not to start a journey of information wonder and discovery. The goal is to get a link or an answer and get out of there. Currently, these LLM-based systems are too slow to compete.
I think the big question is less about the technology and more about the product. Everyone, including Google, believes that AI will help search engines understand questions and process information better. That’s the norm in the industry at this point. But can Google reinvent its results pages, business models, and how information is presented, summarized, and surfaced faster than AI companies can turn their chatbots into more complex, multifaceted tools? Ten blue links aren’t the answer to your search, but neither are text boxes for everything. Search is everything and everything is search. It will take more than chatbots to defeat Google.
