The Future of Search Engines: Personalization, AI, Voice, and Visuals

We've become so accustomed to searching the internet—from answering questions and seeking information (informational queries) to purchasing goods and services (commercial queries)—that we can hardly imagine life without this convenience, and the verb "to google" has become a fully-fledged part of our vocabulary.

Moreover, the widespread use of smartphones with constant fast connection makes it possible to do this anywhere (almost) and anytime.

AI has added to this over the last couple of years, making it even easier to find answers.

We've become so accustomed to them that we don't even notice how they're developing right under our noses: how they're changing search results, how they're speeding up, how the results are becoming more and more relevant... and how they're tracking us for the sake of personalization.

Yes, they're tracking—but don't be paranoid. Search engines aren't tracking Ivan Ivanovich Ivanenko personally, but Google.ID 123456 (20 digits), and they're not interested in where you've been—they're interested in whether you've been to a fishing supply store, for example—perhaps you'd be interested in these fishing hooks?

This is all well and good, but what does the future hold?

In short: the very essence of search won't change, and AI won't replace it, but only complement it. And search will become more relevant (matching queries), more personalized, and faster.

Speed

No, search engines won't instantly start crawling billions of pages—at least not until a working quantum computer is invented—and in the future, they will pre-index websites—gradually but relentlessly crawling them and putting them into a huge database called the "Google index" to give you a ready-made answer.

The speed will increase due to the increased speed (and most importantly, reliability) of the connection, the processing speed on the search engine servers, the use of cloud services, and the speed of your device – smartphone or laptop.

Personalization

We don't just need answers to our queries – we need answers that are most relevant to our needs.

To do this, search engine algorithms need to know our habits, preferences, prior searches, and our reactions to them—whether we were satisfied with the results: for example, how long we spent on the site the search engine suggested as an answer to our query.

All this is already happening, and the amount of information about us (with a note, see above) is constantly increasing with each new request we make or just a click on an advertising banner.

What about the future?

First and foremost, this is the increase in the number of "sensors" collecting information about us. For example, the emergence and development of the IoT (Internet of Things) means that your smart refrigerator, simply put, will provide search engines with data on the range and quantity of food you consume, and your smart vacuum cleaner will provide you with the brand of sneakers you wear this morning (and their condition—"maybe it's time to upgrade?").

Secondly, a more precise "cohort" of our actions. We don't know exactly (no one knows, only the developers, and it's a big secret!) how search engine algorithms work. But they clearly require clarification, because just because we're interested in, say, "screwdriver," that doesn't mean we're definitely going to buy it (though we should still see an ad—just in case).

But if we're simultaneously looking for screws and the material they can be screwed into, and physically visiting hardware stores, while simultaneously searching for a YouTube video on "how to assemble furniture yourself," then the combination of all this information will allow us to conclude that we do need a screwdriver, and then we can "load it up" (here, by the way, it's advisable to stop the advertising, because we're unlikely to need two screwdrivers).

That is: more accurate and more comprehensive information about us will allow us to provide better offers.

Form a specific one, not by algorithms, based on your visits to websites or services, not from past purchases (how infuriating it is when advertising offers me the same thing, but cheaper, 5 minutes after a purchase!) - but actually from your “digital profile”, your complete “digital fingerprint”.

The second point: we understand that a search engine has long been more than just an answer to a question by issuing a list of sites.

No, here we reach you based on your preferences: video views, route suggestions, selected products that might interest you (they probably will!), and even answers to your questions—it's a huge, planet-scale complex that works toward one goal: to give you the best, most personalized information.

The more data the system collects, the more complete your "digital profile" becomes, and the more accurate the personalization, the better and more relevant the information offered will be.

AI

Most of us think the AI ​​boom is a recent phenomenon. But that's not true: there's a Cthulhu-like ancestor of all AI: the search engine algorithm.

By the way, why shouldn't you create texts with the help of—or rather, not "with the help of," but under the "full authorship" of AI—Google will easily recognize them, and the indexing of a site with such texts will cease to exist.

So, if you hear that it's no longer necessary to optimize websites for search engines, as if everything is now done through AI – and even a new industry has been invented called AEO (AI Engine Optimization) – then this is not entirely true, let's say – or more accurately, not the whole truth.

AI algorithms operate on "roughly the same" (remember, no one knows exactly how) principles as "old" search engines—that is, in order to offer your website (and therefore your commercial offer) to the end consumer, they must first find and index the website that contains your offer.

See the "Speed" section above – instantly finding the right answer among billions of websites is physically impossible now and will remain impossible in the near future. Therefore, in addition to "building a language model," AI also pre-searches for websites that contain the best answer and then suggests them.

Voice search

Yes, it still exists: Voice search is supported by over 8 billion devices, and in the US, approximately 40% of users use it regularly, while worldwide, it's about 25% (in Ukraine, 22%).

This difference is probably due to the fact that search engines are initially “sharpened” for the English language, and foreign-language users have problems clearly recognizing what is being said.

The thing is, unlike text (typed) searches, voice queries are more natural. While we might type keywords into our search, we might say something like "buy delicious pizza in Kyiv, inexpensive, with delivery" (so-called "keyword jargon"), we're more likely to say "where can I buy delicious pizza in Kyiv with delivery, and inexpensively."

It turns out:

• longer query;
• text that is more difficult for a search engine to understand.

Clearly, the future of search engines in this area lies in more accurately recognizing what is being said, taking into account local nuances (remember the video about Eleven in a Scottish elevator?), imperfect pronunciation, extraneous sounds, and confusing wording – that is, real “living language.”

Moreover, it is obvious that as recognition improves, users, becoming accustomed and relaxed, will formulate their queries even less precisely.

All of this will need to be processed and provided for – in particular, taking into account personalization and using AI.

Visual search

Nowadays, visual search is represented, for example, by Google Lens – search by image (probably many have used it to recognize “where this photo was taken”), or the built-in camera in Google Translate, which makes it possible to translate (if this can be considered “search” – but now we will explain why we mentioned it) printed texts and inscriptions.

Attempts to launch “augmented reality (AR)” (not to be confused with “virtual reality (VR)”) are already underway, but they have not yet achieved widespread use – for example, smartphones – although it is clear that this is the path by which personal digitalization in the broad sense will develop.

There's a lot we can talk about about augmented reality, but we're interested in "search" as a field.

That is, a wearable device (glasses) for augmented reality should, in our opinion, simultaneously perform three main functions:

• see objects that fall into our field of vision and collect information about them;
• provide information on these objects – in general or upon request (!);
• support voice control and/or Hand Tracking – tracking hand and finger movements as gestures.

In other words, personalization, AI, visuals, voice, and the Internet of Things are all combined into a single system – and… welcome to the future!

Anton Dvorak