How Google changed Search with AI | Mentor Sandy explains | Billion Hopes

Añadido: 2026-05-28

[00:00:04]Hello dear friends, learners, students and professionals. Welcome to Billion Hopes AI. This is Mentor Sandi bringing you AI for real impact. Google truly has changed search comprehensively.

[00:00:16]And it's not just the AI overviews that many people tend to get irritated with, especially the publishers and the creators of websites, but in addition to AI overviews, which are becoming more powerful with each day, you also have the AI mode, the image and video search, the coding within search, and the agentic search as well. So, it's a whole bouquet of services using AI that Google now has baked into search itself.

[00:00:46]And because search is used by several billion people on a regular basis, the entire human experience of interacting with information and the internet is undergoing a fundamental transformation because of Google. You know, I'm reminded around 18 months ago, many people were writing off Google that in the AI race, just like Apple couldn't make it, even Google seems to not have really made it. But then things changed, and in 2025 when the Gemini latest Nano Banana Pro and the Gemini Pro etc. models were launched, suddenly Google leaped forward. So, never ever bet against Google given that it is one of the finest and the first actual data company in the world holding humanity's entire data well indexed. So, today we are going to see how Google changed search with AI because this is what all of us need to know as we use search every day multiple number of time. So, shall we begin the session today? Be with me till the end and enjoy the session to the fullest.

[00:01:48]>> [music] >> So, my dear friends, Google has now built a whole new world of search that brings AI for everyone.

[00:02:05]In the old search, we had just the links. And that is what we did for more than 20 years, all of us. We would use Google, we would get multiple links. In fact, we would never go beyond the first page, 10 links. Hardly anyone would ever click on page two or three, and we would be happy. We would find something of interest, we would go to the website that the link would direct us to, spend our time reading it, extracting stuff, creating what we needed, and it was a painstaking, laborious way of doing things. And literally hundreds of millions of people use search that way for decades, for at least two decades.

[00:02:42]And then, in November 2022, we had OpenAI's ChatGPT launch.

[00:02:48]So, Google was under a mortal threat because people might have stopped using search. Now, that is completely out of the way. People haven't stopped using search. In fact, more and more people are now using search given these AI features. So, now, right now, in the month of June, July, and August 2026, you can use Google Search to answer, discuss, see, act, even code directly from search. And then, of course, personalization of search, which is the next big leap they are taking.

[00:03:22]I understand that each stage, there are very serious concerns about privacy, about the way we are surrendering ourselves to these companies, and so on.

[00:03:31]This way or that, if we are using AI, then yes, these are definitely the downsides of it.

[00:03:39]Absolutely, I agree.

[00:03:41]One by one, I would wish to I want to take you through all these applications.

[00:03:45]And before that, this is the old search. You had a search box, that beautiful search box, that minimalist search box, which actually way back in '98 gave Google a head start over Lycos and Ask Jeeves. They were all full of banners. Google said, "Nothing doing with the banners. The home page will be neat, clean, tidy, white." And that's how the world took to Google. Of course, they had a wonderful back end, the search engine actually.

[00:04:11]And within the URL, you could on page one Actually, no one really went to page two or three usually.

[00:04:18]But times have changed.

[00:04:20]Now you have AI overviews, AI mode, visual search, search agents.

[00:04:27]These are the top four new applications that you can directly use from search itself.

[00:04:36]Most people are already aware of AI overviews because Google throws it by default. All of us get AI overviews by default. And a lot of old users of search skip it and go down.

[00:04:53]But a lot of young people they might actually get hooked on to just the AI overviews and not even look at the links below.

[00:05:03]If you go by the feedback in social media, mainstream media, anywhere of leading publishing houses they are all crying about it. In fact, it's been more than a year that most of the website creators and publishers are claiming that their traffic has gone down by 50% or even more in some cases.

[00:05:26]Many publishers of absolutely specific genre and specific niches of sites are saying that our entire content has been scraped by the AI engines and at times given the right prompt they regurgitated verbatim.

[00:05:41]So yes, this is unethical because no one was compensated for it. But that is the biggest question in AI now. Who compensates the world for what AI companies are earning because they scraped the wisdom of humanity and they are not sharing it with anyone.

[00:05:54]So, before I dive into this Google practical things in search and AI, it will do us a lot of good to remember that in the coming years maybe these windfall gains that the AI companies, new and old, big tech companies like Google might potentially make from AI, there might be specific taxes on these companies to share the gains with a wider world. But, that will happen in time.

[00:06:21]So, let's start with AI overviews. This search generated experience was how it all started a year ago and now it is mainstream. Then AI mode, then visual and multimodal search, then search agents and agentic search and finally agentic coding itself and then the last level is personalization.

[00:06:37]So, one by one, be with me till the end.

[00:06:39]Since you're here, I request you sincerely to please subscribe to my channel Billion Hopes AI and there's a wonderful button join. There are three levels of membership on the channel.

[00:06:48]You'll definitely find something interesting. Please go ahead and join our membership as well. Thank you.

[00:06:54]Dear friends, AI in Google search, this is one. AI overviews, this is how the whole story started.

[00:07:02]They called it search generative experience initially and then it actually became AI overviews.

[00:07:09]Now, Google gives a very quick multi-perspective AI generated answer inside traditional search results. So, this was a huge risk Google was taking because its entire advertising business runs on links and the various advertisement that come along with it, etc. The links itself are advertisements and then you have AdSense, AdWords, the whole thing.

[00:07:32]Literally hundreds of billions of dollars worth of business annually.

[00:07:36]But, the AI overview could have cannibalized a lot of it, but it hasn't.

[00:07:40]In fact, everything is growing.

[00:07:42]So, Google gives a very quick and multi-perspective AI-generated answer right there. We'll do it practically right now.

[00:07:48]And the latest augmentation is it's available across the world for most queries.

[00:07:53]And the inline citations have improved and even community perspectives are being taken.

[00:08:00]So, let's try it out with explain black holes. Let us see what Google gives us.

[00:08:08]Okay, so here we are.

[00:08:10]And the question I want to now put in is explain black holes and I have done it.

[00:08:16]So, now let us see what it does to us.

[00:08:19]It does to our query. Explain black holes, that's it.

[00:08:24]Now Google says, "Okay, I have explained it. This is AI overview." As you can clearly see, this AI overview comes with that Gemini logo watermark.

[00:08:36]A black hole, blah blah blah, and here are the inline citations.

[00:08:41]So, they try to give authoritative citations so that they look better.

[00:08:45]Key anatomy, how do they form? So far we haven't gone to the links and a small YouTube video.

[00:08:53]So, this is how the AI overview works by default. Every user get it no matter what you search. Now, if your search query is very very specific, they may not have an AI overview because they may not have enough data to create an AI overview. AI ultimately depends on data.

[00:09:07]No data, no AI.

[00:09:09]So, now here you can see very clearly the links now start from here.

[00:09:16]All the links start from here.

[00:09:18]Earlier they used to have this people also ask. It's still there and the links come after this. Videos and etc. etc. etc. And of course many many pages of this. So, this is how this AI overview is available to everyone. Every user automatically gets it and this is what has really whacked the business model of publishers and websites. I hope you're enjoying the video. If yes, please do like it and do leave a comment regarding these experiences that you have had with Google.

[00:09:47]So, let me now share with you five more potential examples that could help you understand the power of AI overviews. What is the difference between AI, machine learning, and deep learning with examples?

[00:10:00]If you feed in this question in Google search, the answer will surprise you. It will be a rather high-quality answer. Imagine trying to do this before the AI overview era.

[00:10:12]You would have taken a lot of time compiling everything from multiple sites. AI does it for you. And that's how society itself is changing and people's learning habits will change and in fact a lot of people may simply do cognitive offloading, which is a very, very dangerous thing. You should never allow AI to take away your ability to think. What are the pros and cons of using AI in education?

[00:10:35]How can a small business use AI? Why is climate change causing more extreme weather? Blah, blah, blah. And what should I learn first if I want to start a career in data science?

[00:10:44]Try out these queries and see how the quality of AI overview changes. And it's a wonderful experience. So, I hope you have enjoyed this first quick exposure to AI overviews. That's something you have been using already or maybe you have been skipping.

[00:11:00]But maybe you should try it.

[00:11:03]Now we come to the second application of AI in search, which is AI mode.

[00:11:08]Well, you wanted a chatbot. I'll give you a chatbot. Here is the chatbot.

[00:11:13]I will just click on AI mode. This is AI mode. I have clicked it.

[00:11:19]And my search page has turned into a chatbot. That's it.

[00:11:26]It has just turned itself into a chatbot.

[00:11:29]And the answer to the black hole thing in a chat back form, chat bot format has been thrown at me.

[00:11:36]I hope this is clear.

[00:11:38]And then it will ask you, how was it?

[00:11:41]Etcetera, etc. Got it? So, this is the AI mode.

[00:11:45]Now, the beauty of the AI mode is you can just continue interacting with it as if it's a proper LLM chatbot interface.

[00:11:55]So, this is what Google did. It brought Gemini. The back end is of course Gemini. Google wouldn't be sending API requests to open AI, I'm certain. So, this Gemini thing is baked right into search here.

[00:12:07]Now, the sheer network of Google, the distribution power of Google, makes this an unbeatable combination. So, this is why betting against Google was always a bad idea.

[00:12:19]Yes, it was slow initially, but the moment it picked up, now it has integrated AI everywhere. Some people are revolted by this very idea. I completely understand. But, what is happening? I'm only commenting on that.

[00:12:31]I hope you understand that.

[00:12:34]So, Google search transforms itself into a chatbot.

[00:12:38]And this is being rolled out worldwide.

[00:12:41]And suppose you say, compare AI, ML, and deep learning for a beginner, and now explain with examples from my daily life, this AI mode will accordingly update, and then you have used chatbots, you know how it is.

[00:12:55]So, my dear friend, this is how the AI mode works. I'm giving you five more very good examples that you should try right now.

[00:13:03]Try typing, for example, question four, explain the Russia-Ukraine war from the beginning, then ask me what level of detail I want next.

[00:13:10]And you will be surprised at the quality of the answer. At times, factually, they get things wrong because ultimately they have no brains. They are just ingesting data and refining it and giving you a refined answer. So, they might go wrong at times. So, this is the second AI mode. but there's much, much more to come.

[00:13:29]The third application of AI in Google is visual and multimodal.

[00:13:34]You can now do your search using image, a screen content, lens, circle to search, or video.

[00:13:42]So, earlier it was only with static images. Now it has moved way beyond with these multiple modes available.

[00:13:52]And you can do it very easily from your mobile phone. For example, you can just fix your camera on the real world, switch on a camera in the mobile phone, and you can just focus on any broken item and you want to repair it, and you can ask, "How do I fix it?" Google will use that running video to diagnose the issue. This is very powerful and actually give you a suggestion. Circle to search also does that. This is how it works.

[00:14:23]Contextual search from any app. You can circle or scribble on screen content and find information without leaving that app.

[00:14:31]So, this is a very deep integration of AI and so-called intelligence right into our day-to-day lives. I hope that is clear. So, suppose you actually are working on a mathematics question and you are stuck at a part, you can circle that complex question and Google will then start giving you detailed options.

[00:14:54]Earlier you might have taken a full picture of it and fed it to open AI or some other chat GPT or Grok or whatever and asked for a solution. Here you can just circle it and it'll start giving you the answer.

[00:15:04]Multimodal search, image and text together you can now use for search.

[00:15:08]Generative AI analysis, you can upload the image for deep understanding that all other chatbots can also do. So, this is a huge upgrade for those who are comfortable using it.

[00:15:20]Then comes agentic search. This is absolute madness.

[00:15:24]You can create agents right from your search bot.

[00:15:30]You can directly manage agents from search.

[00:15:36]So far, all of us for 20 years use Google as an information engine.

[00:15:42]Now Google says, "Would you want to use me as an action engine?

[00:15:45]I can do multi-step agentic actions for you."

[00:15:51]Gemini's bot, a personal agent which will operate in the background, and then also managed agents. All these were launched at the IO 2026.

[00:16:04]So let's try it out, and let us see whether it really works the way it claims it does.

[00:16:09]So here's the prompt I have actually my need. Find me a good restaurant for six people in Saturday evening near Connaught Place.

[00:16:17]Compare the options, check my calendar, and book a table. My God, so I'm going to do all this and right from search.

[00:16:23]Remember, you need to use AI mode for this, okay?

[00:16:28]Just put it here.

[00:16:31]And I put my question here. In fact, let me put it in the text format.

[00:16:36]So I have pasted it in the text format now, and click.

[00:16:41]So I've asked it to find me a good restaurant for six people this evening, Saturday evening, near Connaught Place, blah blah blah. All my requirements have been specified. The more specific, the better would be the answer. It has started.

[00:16:53]It has started giving me multiple options, and then it'll give me all the web links, all the phone numbers, etc. etc. And I can then start the agentic action with it.

[00:17:14]So my top recommendations are here, and it has given me various time slots. So, suppose I click this.

[00:17:21]Then reserve with Google, the action commences here. So, this is Agentic AI.

[00:17:26]What would you want to continue with?

[00:17:28]And everything has started here.

[00:17:31]So, this is actually Agentic AI.

[00:17:33]You can go deeper into it and actually try and complete an action and see for yourself. That's not the goal for today's session, but I just wanted to tell you that Agentic search is now built straight into the AI mode and it is better you start with a new window each time you have such a request. I hope that is clear.

[00:17:53]So, here are some examples of what you can do. You can do restaurant booking, travel planning and booking, event ticket search, appointment booking, shopping assistant, everything you can complete it from there. And if you have been making agents separately, you can use them as well. But Google, sensing that agents for day-to-day work are now a big thing, is taking coding away from it.

[00:18:17]So far, you would have needed some coding mind to do agents, to build agents. Google says, "Why do that? Just come to me, I'll do it for you."

[00:18:27]Of course, there are many limitations of this, but for an absolute newbie, this is a godsend.

[00:18:33]Then Agentic coding, this is the future.

[00:18:37]What it is saying is search, of course, builds the ideal response.

[00:18:41]Now, if you want a custom generative user interface also, you can do it straight from search.

[00:18:53]Got it?

[00:18:54]So, for example, you want to have a new health and wellness routine, you can ask search to create a health and wellness tracker. And it'll give you multiple options and will start building a UI for you.

[00:19:08]This actually defies imagination. This was something that many startups would have actually relied for on their revenue model.

[00:19:18]But now the power of Google anti-gravity and agentic coding and Gemini 3.5 flash all three are coming into search itself.

[00:19:29]The idea is very simple. Google wants to capture this market and create products that reach billions. At the same time open AI and Anthropic are building frontier models and better research and more cutting edge. Google is doing all that, but it is now very aggressively pushing these products combining everything and presenting inside search itself. So, I think that might be a game-changer in time.

[00:19:54]And last but not the least personal intelligence in search.

[00:20:00]This actually is a cutting edge thing, but a very dangerous thing.

[00:20:04]What Google says is if you allow me to connect everything, Gmail, Google Photos and all other I will give you very customized solutions to everything because that's your personal intelligence.

[00:20:20]Shopping recommendations will be tailored.

[00:20:22]Your personal health issues will be tailored or your technological issues will be tailored.

[00:20:28]And I know exactly how you live because you have allowed me to track that so I can give you suggestions for your food, your daily schedule, etc. Long story short Google is determined to change search with AI or rather bake AI into search because search is the widest possible product it could ever have created, reach product, and well, it actually has brought AI.

[00:20:56]Whether people accept it and start using it is another matter altogether, which we shall see in due course of time.

[00:21:04]What kind of a challenge with Perplexity like browsers or Perplexity like online search services that connect multiple LLMs pose, we will see that in our next videos.

[00:21:14]I hope you have enjoyed it. Please do like this video. Please do subscribe to our channel Billion Ops AI. This is your mentor Sandy bringing you AI for real impact. See you soon.

[00:21:28]>> [music]