Just a few short months ago, Google found itself facing a competitor like it had never faced before. A young upstart, one with a powerful blow, was circling, threatening the search engine’s supremacy as the long-reigning, undefeated, undisputed heavyweight champion of online search. That may sound like hyperbole. But for the first time in two and a half decades, the search engine landscape was changing.
It was February. ChatGPT had just set the record for the fastest-growing user base, reaching 100 million monthly active users in just two months. The news that Microsoft was to integrate the AI language processing tool into its search engine infrastructure signalled a once-in-a-generation shift in search behaviour and ushered in what many industry insiders believed to be the search AI wars.
Microsoft, the perennial runner-up in consumer search engine usage, saw ChatGPT as their opportunity to claw back a percentage market share from Google. But alas, they were thwarted. Sure, Bing was the first to incorporate the OpenAI platform, GPT-4, into search earlier this year; however, that didn’t see users abandon Google and flock to Bing.
And now, Google is on the brink of pulling a rabbit out of a hat. One with the potential to forever change the search landscape.
So, just what did Google do? Announce that they intend to supercharge search with generative AI.
A New Search Engine: Google Search Generative Experience
Forget algorithm updates. Google Ads. Forget wearable technology. Google Nest. Google Generative Experience is the biggest evolution to Google’s core service or product offerings since the launch of the search engine itself.
Google has yet to fully roll out a search ‘experience’ like no other (for that, we’ll have to wait for the Core Web Vitals update in March 2024) * however, from the short snippets, GIFs, and other demonstrations, it’s clear that we’re on the cusp of a revolution.
But what exactly is Search Generative Experience (SGE)? Let’s briefly examine the technology that will likely change how we experience search engines.
SGE won’t be a chatbot on steroids. It won’t be a redesigned search experience featuring a solitary messaging-style search bar for users to type queries into. Google is putting AI front and centre on the most valuable real estate online: its search results.
Users will type a query into Google as usual. Let’s say, “Who won the World Cup in 2022?” then hit enter. Google will load its standard search results. The World Cup winner will appear. What users searched for. Similar users also asked. Links relevant to the search query. But there’ll be something new.
Immediately above Google’s normal search results, another section will appear. One with the message, “Generative AI is experimental.” Wait a few seconds, and an AI-generated summary will appear. One that’ll display a few paragraphs outlining the winners, the circumstances of the game, maybe what the victory meant to the Argentine people, or who Argentina beat in the final and the score.
Google calls this the “AI snapshot.” The information displayed has been collected from sources online and displayed by Google’s LLM.
Users can expect AI-driven LLM results in March 2024.
Have a conversation with Google
Some of the more expectant users might, at this point, be asking, “Is that it?” Well, no. In addition to the answers generated by prompts, Google SGE will return website links within its clickable AI-generated content results that users can click for more information. The generative AI answer box will change colour according to query intent and journey.
But there’s more.
In addition to pulling results from Google Shopping to help users find a specific product or service, Google SGE allows users to have a follow-up conversation or ask for additional prompts to discover further information.
Users can continue a natural conversation to explore a topic. Google SGE will recognise context and perspectives to further tailor interactions, something that can be especially helpful to users who wish to ask more questions or seek out complex information.
The AI learns user intent by recognising context and then reformulating subsequent interactions to better understand and meet user intent.
In simple terms, it’s less like users are firing questions at Google and more like a conversation.
The way people search will change. It’s just a matter of when
At the moment, the public’s stance on Google SGE is that it’s just an experiment. One that’s being tested by Google Search Labs, an aptly named new subdivision of Google Labs. Anyone living in the US and over the age of eighteen can apply to be part of Google Labs. As you’re reading this, it’s likely that someone in the US will be testing Google SGE!
As of May 2023, Google’s public stance is that its AI search integration is an experiment. However, it’s also clear that the foundations are being laid to change the way users search and interact with Google online. In several ways.
Think about how we use Google to buy products and services. For example, you want a bike. You search for a bike. You find any retailer. You intend to buy a bike. But what type of bike? What about specifications? Price? Shipping? Retailer?
Shop with generative AI, and you’ll get a snapshot of all the noteworthy features that influence your decision. From price comparisons and retailer ratings to up-to-date reviews and product images, you won’t need to go hunting online for specific information. Why? It’ll be built on Google’s Shopping Graph, a real-time machine learning AI designed with one goal in mind: to deliver the very best search results.
And that’s just a single feature.
Why is this so important?
OpenAI and LLM (language learning models) aren’t a novelty. In fact, GPT-4, ChatGPT’s most advanced system, is capable of solving problems with a greater degree of accuracy than was possible just three months ago thanks to its broader general knowledge and problem-solving abilities.
More creative and collaborative than ever before, CPT-4 can generate, edit, and iterate with users on a seemingly boundless number of creative and technical writing tasks, including composing songs, writing screenplays, or learning and repeating writing styles.
Combine this with OpenAI’s reasoning capabilities, which afford users the chance to streamline and simplify information by leveraging greater computation to deliver increasingly capable and sophisticated language models, and it’s not hard to understand why consumers now trust OpenAI to meet a wide range of search queries.
Don’t believe us? ChatGPT receives more than 65 million visits per day. Microsoft Bing, a search engine with a 14-year legacy, gets just 40 million visitors. * ChatGPT has a bigger user base than Bing.
No matter how you spin it, OpenAI platforms are and will continue to take market share from search engines, including Google, unless the search engine finds a way to leverage LLMs to broaden and improve its own search engine services.
And that’s exactly what Google is doing.
Will Google win the future of search?
Now, as we’ve already mentioned, it’ll be 2024 at the earliest before we begin to see AI being leveraged by Google search to truly bolster the user experience; however, when it’s eventually rolled out, it would be naive to assume that it won’t compete with any other OpenAI platforms.
AI has seen a drastic evolution in just a short space of time. AI’s current predicament is not how to get it to work or whether anyone will be interested; it’s how to package the product design and market the technology as necessary to everyone in the world.
This is something that Google has a lot of experience with. We’re at a tipping point for AI, and Google is laying its chips on the table.
Remember, before Google, there was Yahoo. When was the last time you typed a search query into Yahoo?
Resources:
https://www.theverge.com/2023/5/12/23721037/google-ai-progress-search-docs-starline-video-calls
