AI-powered Search Generative Experience unveils the possible future of search

Instead, Google launched an experimental project called Search Generative Experience (SGE). For now, Google is only giving access to users in the US, but it offered up a few examples of what we can expect.

What will the future of AI search look like?

Google gave us a preview of what the future of AI search could look like in a video during its keynote presentation at I/O.

Based on the keynote presentation and video, there are four key takeaways.

1. Generative AI responses for information searches

This feature will probably get the most attention as SGE progresses. It’s the most ChatGPT-like implementation of AI technology that Google teased during its presentation. For relevant queries, Google will generate AI responses that pull information from multiple sources, helping it provide comprehensive answers to more complex questions.

During the keynote speech at I/O 2023, Cathy Edwards offered this example query: ‘What’s better for a family with kids under 3 and a dog, Bryce Canyon, or Arches’.

That’s a lot of information requested from a single query and it’s unlikely that one single page will provide everything the user is looking for. So, instead of simply returning a list of results in blue links, Google can use generative AI to pull information from multiple sources and compile it into a complete answer.

Google will also cite the web pages it gets this information from, allowing users to visit relevant web pages for more information.

2. Answering complex questions in a single session

Generative AI will also help Google’s natural language processing technology show its full capabilities. As things stand, the algorithm can understand pretty complex queries, but it can only return web pages that exist separately.

If we take the same query and type it into Google, the algorithm understands the query, but it’s limited to returning results for individual parts of the query. The more complex queries become, the less likely it is that a single web page will provide all the information users are looking for.

In many cases, users will have to break their query up into multiple parts and piece together the individual bits of information they’re looking for. With generative AI responses, Google can compile this information for users and answer complex questions in a single session.

3. Exploring topics in greater depth

Generative AI doesn’t mean Google will answer every query users throw at it on the first attempt. Even with greater AI capabilities, the search engine is still limited to using the information (data) it has access to.

However, Google is implementing a linear search experience that remembers multiple inputs. So, instead of every query starting a new session, users can explore topics in more detail with follow-up prompts.

“Context will be carried over from question to question, to help you more naturally continue your exploration. You’ll also find helpful jumping-off points to web content and a range of perspectives that you can dig into.” – Supercharging Search With Generative AI

For example, users might start with a relatively simple query like ‘best home cinema systems’. After viewing the initial results, the user might decide they don’t want to buy a new TV but want to improve the audio experience. So, they could follow up with a query like 'show me the best sound systems for watching films at home'. Next, they might follow up with another query like 'show me systems compatible with the Sony X95K TV'.

As users type follow-up prompts, Google remembers their previous queries and delivers more relevant results with each interaction.

4. Helping consumers buy products with generative AI

Google also demonstrated how generative AI will help consumers buy the right products. When users search for a product category like eBikes, Google can return product recommendations with a generated product description and key things to consider when making a purchase.

If users input a more specific query like ‘eBikes in red for five-mile commute with hills’, Google will pull out key details from the query. It will recommend eBikes with enough battery capacity to complete the journey and enough power to tackle inclines.

Google is taking its time with AI search

Instead of rushing out releases to compete with ChatGPT and Bing’s new AI-powered search engine, Google is taking its time.

The search giant hasn’t launched anything new or given us any release dates. It’s testing AI features with its SGE experiment and the approach seems to be enhancing the current search experience, rather than overhauling or replacing it.

Time will tell if this approach pays off.