A few days ago, Google announced plans to dramatically change search by incorporating artificial intelligence into the results. Shortly before the announcement, the company showed off Bart, Google’s AI chatbot. The company has relied on a machine-learning tool they call RankBrain for years as part of the search algorithm.
Is AI Ready for Search?
Google may have rushed the statement in response to an increased focus on ChatGPT. Many people have already signed up to use the free version of ChatGPT from OpenAI. A few days before, Microsoft previously announced that it planned to incorporate ChatGPT technology into its Bing search results, which probably sparked some competition.
Perhaps Google moved too fast. According to Reuters, Bart misspoke in an advertisement, causing the company’s shares to plummet nine per cent. The setback will undoubtedly prove temporary, but it still generated plenty of embarrassing press. Google quickly responded that the error highlighted the need for the rigorous testing process it already has in place.
The issue seems worth mentioning because it underscores one problem with ChatGPT that users have already noticed. Sometimes, the ChatGPT app also offers inaccurate information that appears made up. This public mistake may encourage everybody to exercise caution before rushing to replace tested solutions and, most of all, people with new and disruptive technology. Still, as a Microsoft employee commented, the race is on.
How Does AI-Powered Search Work?
Despite setbacks and criticism, the future has already arrived. SEOs should prepare for traditional search to rapidly evolve into a new experience that will probably rely on some combination of the existing algorithm, human quality assurance, and AI.
To prepare themselves, SEOs and other stakeholders can take some proactive steps:
Understand What Artificial Intelligence Means
AI refers to multiple technologies. These include machine learning, natural language processing, and other disciplines. More important than the meaning of AI, website owners and SEOs should take time to understand what AI should do and why it can enhance search.
AI should help people by making tasks easier, more reliable, and safer. Think of querying a search engine and viewing a list of website links that could offer a solution. The current process often involves several steps:
- The user might repeatedly need to refine the query because the first try didn’t yield acceptable results.
- Finally, the search user can start checking websites in the listings to see if they answer the question.
- Sometimes, users need to open another window for more details or clarification on an initial result.
Artificial intelligence strives to better discern the user’s intent by providing a summarized answer and not just a list of web pages. The response may also include sources for the response. For instance, a query about diagnosing a cold could also have references to learn how to treat symptoms, find a local clinic, or have medicine delivered.
Thus, AI search can save time and provide a better experience because the user doesn’t need to perform multiple queries. Unsurprisingly, early demos for Bing’s AI search also included paid ads.
Explore the New Bing to See AI Search in Action
Seeing AI search in action offers the best way to understand how it works and differs from how search engines have worked for decades. Today’s search engine optimization focuses the most on Google. However, Bing might offer the best current examples right now.
Microsoft announced a new AI search that relies on an upgraded version of ChatGPT. Compared to ChatGPT’s public version on OpenAI, ChatGPT 3.5 can retrieve recent information, supply citations, and process queries more efficiently.
The public can learn more by visiting The New Bing. This website offers examples and more information. The site also lets the public join a waitlist for early access. Primarily, AI-powered Bing differentiates itself from traditional search by providing more complete answers to queries instead of just offering a list of links that may provide the solution.
For instance, users could ask for a complete trip itinerary or the details of a five-course vegetarian meal instead of asking for each piece of these complex queries. Even more, the AI remembers the conversation during a session, so it’s possible to ask related questions to refine the answer. Of course, the New Bing also provides source citations, so users can navigate directly to the resources that the AI uses.
Learn How AI-Powered Search Differs From Regular Search
The first internet search engines debuted in the 1990s. Thus, people have had three decades to study them. Qualified SEOs understand a lot. Still, they should admit they can’t know every detail about how search engine algorithms work, mainly because search engine companies keep many details of their algorithms secret.
Determining the best way to optimize for AI search starts with understanding traditional search and how the new technologies compare. Thus, begin by learning what Bing says about the differences between its new AI search and traditional search:
- Traditional search engines offer lists of relevant links in response to queries. The new search result provides links but summarizes the top results into a comprehensive answer.
- The new search can parse complex questions and can better determine user intent. It can also remember previous responses as users ask additional questions to help refine the results.
- AI search also works as a creative tool. For instance, chatbots can write songs, poems, and stories and help generate ideas.
How Should SEOs Optimize Content for AI Search?
Think of AI search as the next step in search evolution but not an entirely new creation. Even though technology has evolved, it will still maintain some characteristics of its predecessor.
Thus, SEOs should still attend to fundamentals like backlinks, relevance, keywords, navigation, titles, and headings. Just as important, pages that load reasonably quickly, function on various devices, and engage the user’s attention enjoy an edge.
Primarily, pay attention to what Google has been telling website owners for years. Craft web pages for users and not for search engines. For instance, consider keyword spam an obsolete tactic. Improved language processing has rendered it unnecessary and counterproductive. Keyword spam can damage readability and rankings, just as it has for some time now.
Focus on Relevance and Not Keywords
Google already employs a machine-learning AI called RankBrain to process queries. RankBrain excels at ferreting out relevance, even when no keyword in the search query exactly matches one found in the content. Thus, write naturally for humans and worry less about taking extraordinary steps to add keywords or phrases for search engine bots.
Consider User Intent
Even today, search engines like Google don’t just consider what users look for when they search. The algorithms also try to understand why the user made the query.
It helps to break search intent down by considering these four kinds of queries:
- Transactional: People often want to do a specific action and need to find a page they neglected to bookmark. User intent could include sending a message or paying a bill.
- Informational: Searchers may want to see general information on a topic without necessarily intending to take additional action. For example, a user might search for the rules for playing a card game or the height of Mt. Everest.
- Commercial: In this case, users want to make a purchase or, at least, comparison shop for something they mean to purchase eventually.
- Navigational: Searchers want to find a specific website or page but can’t remember the URL, so they turn to search.
The User’s Next Intent
Determining why users will search for a specific topic can present a challenge. However, pages likely to rank higher with AI search for some queries also take one more step. These pages provide the user with resources for their subsequent possible actions.
Trying to predict the user’s next action even helps rank informational pages that users reach with no clear goal in mind. Consider how heavily-interlinked sites like Wikipedia perform well in search with the “height of Mt. Everest” query used as an example. The Mt. Everest Wikipedia page offers links for additional information, like people who died on the mountain and which countries the mountain rests on.
A 30-second query for information can quickly turn into hours of following rabbit holes or, in the case of an AI search, an extended conversation with the chatbot, thus demonstrating engagement. For websites designed to rank for queries with commercial intent, consider the buyer’s journey that may progress from curiosity to comparison shopping to purchasing.
Other Factors Signaling Intent
Search engines like Google have reams of historical data they can use to narrow down user intent. Sometimes, the search engine will provide different results for similar searches on various topics. For instance, a search for “Chinese food” typically focuses on nearby restaurants and food delivery. Lower down, the results page will have some suggestions about types of Chinese food or recipes.
Other factors that an intelligent search engine uses to determine intent include the time of day, location, and the kind of device. For instance, mobile searches for Chinese food may indicate the user wants driving directions to a nearby restaurant. Still, a desktop or smart device user could prefer recipes or menu suggestions.
How to Optimize Websites for AI
Most importantly, content producers and SEOs shouldn’t try to think like traditional search engine bots. They need to think like people, which is what artificial intelligence attempts to do. Search results will include more content about a topic and not just a list of options, so relevance and search intent become even more critical ranking factors. Also, gaining visibility should prove more competitive as the results favour offering more content instead of several links per page.
Next Steps for SEOs and Optimized Content Producers
More concerning, AI-enhanced algorithms may start devising their own ranking factors. Google announced that machine learning had already begun developing its own rules as long as five years ago. Even more, they may do this inside a black box that few can penetrate and nobody can predict.
The uncertainty underscores the importance of exploring AI search as soon as possible because the best predictions will stem from real-time experiments and testing. Also, some AI tools already exist that may help SEOs. When artificial intelligence takes over search, SEOs will probably need to arm themselves with their own AI tools.