In our time spent online, Google has always been the middleman – a bridge between people asking questions and websites providing answers. For digital marketers, the goal of modern SEO is to work with Google, providing the best answers to relevant user queries. And the reward? Attracting hordes of website users, boosting traffic, and, ideally, turning clicks into customers.
But what happens when Google goes direct-to-consumer? When instead of leading users to your website, they provide AI-generated answers directly in their search results page. If AI is doing the heavy lifting, will websites and blogs still have a place online?
In this post, we’ll take a look back at the evolution of Google Search. We’ll chat about Chat GPT and artificial intelligence, and how generative AI is having its moment. We’ll dive into the integration of chatbots into search engines and end with some thoughts on what this could mean for website traffic in the future.
The evolution of Google Search
Google’s transition from a basic search engine to an omnipresent information source has been a remarkable, AI-driven journey.
Starting with PageRank, its pioneering algorithm introduced in the late 1990s, Google quickly became synonymous with search (Ask Jeeves, who?). This system ranked web pages based on their relevance and popularity, laying the groundwork for highly efficient search results.
Smaller updates were introduced until a pivotal shift in 2012, with the introduction of the Knowledge Graph. This allowed users to obtain instant answers to queries, presented directly on the search results page. Though it bypassed the need for users to visit specific websites, unlike generative AI, results were still pulled in directly and dependent on external website copy.
RankBrain’s integration in 2015 was another pivotal moment, where AI and machine learning was leveraged to comprehend the context behind queries, improving search relevance. These steps transformed Google Search from a keyword-driven, lookup tool to an intuitive guide through the expanse of information online.
So it’s important to remember that, while AI is buzzworthy, it’s not new to Search. This evolution has been a stepwise process through countless algorithm updates. And while generative AI presents some uncertainty for webmasters, Google continues to move towards the same target – an efficient and seamless user experience.
The next iteration of Search
Google has become an expert at predicting what users are looking for, crawling unthinkable amounts of data to find it, and delivering the best solutions – all in what seems like an instant. However, as a business, there’s significant risk in sending users away from their platform. Google makes a lot of money from advertising, and it’s in their best interest to keep users coming back and increasing the time they spend in search engine results pages (SERPs).
Knowledge Graph was a first-step solution in keeping users on-site, pulling answers from websites directly into search. It also facilitated voice search, providing solutions to users in the most convenient way possible.
But those solutions still relied on other websites and sourcing information externally – the next step was moving in-house. Doing so opens up a new horizon for search, and it comes in the form of Generative AI.
Generative AI: making something from nothing
Generative AI refers to a category of artificial intelligence that focuses on creating or generating new content, data, or information, rather than just analyzing or processing existing information. It’s capable of producing original content, such as images, text, audio, or videos, that closely resemble human-created content.
These AI models, often based on deep learning techniques like neural networks, are trained on vast datasets to learn patterns, correlations, and structures within the data. Generative AI models can then generate new content by predicting and synthesizing information based on the patterns they’ve learned during training.
Open AI & ChatGPT
One popular example of generative AI is OpenAI’s GPT (Generative Pre-trained Transformer) series, including models like GPT-3. These models can generate human-like text based on prompts given to them, demonstrating the ability to create coherent and contextually relevant content.
Generative AI has applications in almost every field, including natural language processing, content creation, image synthesis, and even drug discovery. Its ability to create new content has sparked interest and innovation in diverse industries, pushing the boundaries of what AI can achieve in creative and problem-solving domains.
Who’s using it?
OpenAI released the GPT-3 API for public access in June 2020, making it available for developers and organizations to utilize for various applications. The popularity of Chat GPT and the GPT-3 API grew steadily following its release, with developers, researchers, and enterprises exploring its functionalities and integrating it into a wide array of platforms.
While already available for a few years, 2023 saw a rapid increase in search interest for the platform. This can be attributed to a few factors, such as availability in more countries, increased integration, and refinement. Changing the online landscape as we know it, a better question might be “who’s not using ChatGPT?”. Looking at Google Trends, we can see a growing search interest over the past 12 months.
Narrowing in on the last 3 months, it’s interesting to note that search interest in ChatGPT is cyclic, consistently dropping on weekends and peaking mid-week. This highlights the fact that many users are incorporating this technology into their workflows.
How are they accessing it?
Early access to Chat GPT or the GPT-3 API was primarily available to developers, researchers, and businesses through OpenAI’s platform. This access typically required applying for API access, which involved a process of registration on OpenAI’s website.
For public-use, OpenAI released an early demo of ChatGPT in late November 2022. According to Forbes, “The chatbot quickly went viral on social media as users shared examples of what it could do. Stories and samples included everything from travel planning to writing fables to code computer programs. Within five days, the chatbot had attracted over one million users.”
Today, most users access the chatbot by visiting chat.openai.com or through the ChatGPT App – both of which require signing up for a free account. This provides access to GPT-3, or users can upgrade to the Plus version (GPT-4) which offers advanced access for $20/month.
As separate entities, ChatGPT acts as a direct competitor to search engines. Seeing the potential, the goal quickly became integration. And in a surprising turn of events, a lesser-known underdog beat Google to the punch.
Bing’s launches Bing Chat, powered by GPT-4
In February 2023, Microsoft’s Bing unveiled its new, AI-powered search engine. Using GPT-4, the latest iteration of GPT from OpenAI, searchers are prompted to use Bing Chat to get concise, AI-generated responses.
Note: SEMrush has a great article on Bing’s AI if you’d like to dig deeper.
After Microsoft’s investment of more than $13 billion in OpenAI, what was originally a slowly-declining platform became a real contender in the search-engine space. In addition to Bing search, Microsoft is integrating this technology into its sales and marketing software, GitHub coding tools, Azure cloud and Microsoft 365 productivity bundle.
With Bing page visits increasing by 15.8% since the launch, this integration has been hugely successful. We’ve also seen a significant rise in the value of Microsoft stock, climbing to an all-time high this December after the proposed hiring of OpenAI’s CEO.
While Google has long dominated the Search market, the infusion of OpenAI’s technology into Bing’s ecosystem triggered a response from Google and other competitors. Looking to further enhance their search algorithms and features, AI integration in Search continues to push boundaries and set the stage for a more competitive landscape.
Google’s AI Chatbot: Bard
Earlier in an October 2019 algorithm update, Google introduced BERT, a natural language processing model. This brought on a new era of search comprehension by understanding context within search queries, including subtle nuances like sarcasm and colloquial language.
Building on this technology, Google launched their AI chatbot Bard in March 2023. Instead of GPT, Bard is powered by Google’s newest language model Palm 2 (Pathways Language Model). Designed to generate human-like responses in a dialogue, this technology has applications in chatbots, virtual assistants, and customer service.
How to access it
The Google Bard API is currently available to developers, researchers, and enterprises. Generative AI models and fully managed tools, like Vertex AI, make it easier for developers to integrate this technology into their own applications.
As a stand-alone platform, Bard is available to the public and offers content in more than 40 languages. If you’re in one of the 180+ countries that currently support it, you can access the chatbot by visiting bard.google.com.
Unfortunately for Canadians, we didn’t make the list, and there’s been no indication on when Bard might become available to us. Along with the EU, Google has hinted at regulatory reasons for this issue.
Much like the Bing chatbot being integrated into Bing’s search engine, it’s anticipated that Bard will be integrated into Google Search in the near future.
Bard vs. ChatGPT
The biggest difference between Bard and ChatGPT is the datasource. The latter pulls from an updated dataset – but we’ll let it explain itself:
Bard, on the other hand, uses real-time information pulled from the internet. This means that for current events, research and time-sensitive information (like weather patterns or airline seat sales), Bard has the advantage.
For content-creation tasks, ChatGPT is said to be superior. If you’re interested in learning more about the differences between the two, Tech Target has a great article on the topic.
Chatbots in Search
Compared to search engines using keywords, AI chatbots use natural language queries for search. Like all large language models (LLMs) tasked with generative AI, responses can be prone to inaccuracies known as hallucinations. Google recently offered a solution with their “Google it” button. This allows users to fact-check responses from Google Search without leaving the Bard platform.
But what about bringing chatbots into search engines? Like we saw earlier, Bing has already integrated their chatbot, and highly successfully at that. Contrary to what was expected, websites have also seen increased traffic from Bing. AI-generated answers still link to relevant websites, which promotes users continuing their search in external websites.
But is the increase in website visitors from Bing simply a result of more users using their search engine? Will we see the same results from Google’s Bard once it’s integrated in search?
Predictions for 2024
Generative-AI’s impact on search engines
Looking ahead to 2024, the trajectory of search seems promising. Anticipated advancements in AI integration, further refinement of algorithms, and a deeper understanding of user intent are poised to transform search experiences.
With generative-AI integration, there’s no doubt that more users will be using these platforms. And while ChatGPT continues to dominate the public-use market, only time will tell which players dominate and which are left behind.
Personally, I don’t think Google is going anywhere, but I’m also very interested in what Microsoft & Bing do next. Fourteen-year-old Carla thought MySpace and MSN Messenger would be around forever, so who knows what the tech future holds.
Bard’s impact on website traffic
One of the big questions we’re facing, as a digital agency, is what happens to websites when users can find everything they need on the results page. For most of our clients, organic traffic makes up the lion’s share of their website traffic. What happens if Google no longer sends users their way?
My prediction is that for 2024, Bard will be used in much the same way as Knowledge Graph. Users will be provided with an instant, higher-level understanding of a topic. But with frequent inaccuracies in responses, I think most users will use it instead as a jumping-off point. They’ll continue to research using external websites, even following the citations and sources that Bard now provides.
Once Bard is well-integrated into Google Search, I anticipate seeing decreases in organic traffic – especially for websites that are information-based. However, I don’t think this will result in decreased customers or conversions. Whether that information is being presented on your website or generated in a SERP, what’s most important is that it is building trust for your business.
I also predict that users who end up on your website will be further down the conversion funnel, and that conversion rates will increase. Users will be closer to making a decision, whether it’s to contact you about a service or to make that purchase. And when that time comes, I predict that they’ll want to go straight to you as the source.
Embracing the winds of change
Is there uncertainty about generative AI in search? Absolutely. I think it’s safe to say the whole world is worried about AI, whether it’s about AI replacing jobs, a lack of regulation or general intelligence surpassing our own.
While we’re not going to answer these global questions in a 2000-word blog post, Google has given us some tools and suggestions to prepare websites to thrive in an AI-search environment. One example is implementing structured data or Schema into your website code. This markup helps search engine crawlers (and chatbots) better understand and pull information from your website, making it easier for them to provide relevant answers.
Why make it easy for them?
One thing we’ve consistently seen with search engines is that the more we help streamline their crawl and provide great user experiences, the more our efforts are rewarded with high-quality website traffic.
The future promises an even more intuitive, personalized search experience. And while the winds of search are shifting, rather than battling against AI-driven currents, we’re better off harnessing them to propel us forward.