Episode Categories:
Resources:
Join the discussion: Got something to say? Let us know on LinkedIn and network with other AI leaders
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Start Here Series in our Inner Circle Community: Join for free access
AI-Powered Search: Transforming Information Retrieval into an Intelligent Operating System
Today’s business environment demands not only rapid access to data but the contextual intelligence to act on it. The latest advancements in AI-powered search, specifically within Google Search and Gemini, are redefining how information is accessed, processed, and personalized, creating a new layer for business efficiency and decision making.
Personalized Search: From Keyword Queries to Contextual Intelligence
Traditional search depended heavily on keywords and manual filtering. Now, AI-driven personal intelligence is changing the underlying mechanics of information retrieval. By securely connecting to individual data sources such as Gmail and Google Photos, AI can contextualize requests and generate answers aligned with the user’s real needs. For instance, when searching for recommendations or making purchasing decisions, the AI draws on recent emails, purchase histories, and even image data to suggest hyper-relevant options, eliminating the need for multi-tab investigations and manual filtering. This self-optimizing search replaces repetitive tasks—like typing "restaurants New York" and scanning dense lists—with a personalized preview that understands travel plans, preferences, and constraints, offering recommendations that are not only accurate but actionable.
Conversational Search: Effortless Exploration and Decision Support
New conversational capabilities within Google Search enable seamless back-and-forth dialogue, supporting complex explorations that mimic real-world decision processes. Users can now ask multi-layered questions, refine their requirements, and receive AI-generated summaries, all within a single open thread, particularly effective for multifaceted business queries such as market research, vendor comparisons, or trip planning. This approach drastically increases user-reported satisfaction, as conversational search minimizes the friction associated with traditional search methods, and supports dynamic follow-ups—business leaders no longer need to start over for every adjustment but can clarify, drill down, and iterate within the same interface.
Agentic AI: Task Automation and Proactive Recommendations
The integration of agentic AI features in search expands its utility beyond passive data retrieval, moving toward agent-driven task automation. In scenarios such as booking reservations, scheduling services, or negotiating deals, Google Search now leverages AI to not only suggest actions but execute them. For instance, the system can analyze a company’s intended trip, find optimal hotels based on budget, location, and individual company preferences, and even initiate bookings—all without direct user intervention. Additionally, AI-driven agents are capable of running background research, contacting local service providers, and delivering consolidated results, thus freeing up valuable business hours and enabling immediate action.
Multimodal Capabilities: Expanding Reach Beyond Text
The deployment of multimodal models—like Gemini 3—enables search tools to interpret and respond to queries spanning text, images, audio, and live data streams. Business applications include on-the-fly analysis of documents, comprehension of visual materials, and real-world object recognition via smart devices. This is particularly useful for companies operating in logistics, retail, or customer experience, where the ability to rapidly process visual and auditory information can translate into real-time operational decisions or customer insights. On Chrome, AI-powered drawer features and contextual assistants enable users to question and receive analysis about any web page, elevating in-browser research from passive reading to interactive, AI-assisted understanding.
Business Value: Practical Efficiency, Enhanced Decision Making, and Contextual Action
The intersection of personal intelligence, conversational and agentic AI, and multimodal functionality presents actionable opportunities for business owners and decision makers:
Time Savings: Routine yet complex tasks—such as research, cross-platform analysis, and filtering—are handled automatically by AI agents, reducing manual effort.
Contextual Value: AI-powered search can proactively identify gaps in planning, suggest optimal choices before deadlines, and adapt output based on individual business context and operational needs.
Integrated Experience: Personal data securely enriches search outcomes, ensuring that recommendations and actions are tailored to specific business requirements, from purchasing decisions to operational planning.
Continuous Learning: Over time, AI models learn user preferences, business routines, and intent, enabling increasingly precise results and recommendations.
Action Steps for Businesses: Maximizing Return on Search AI
To harness these advancements, businesses should:
Encourage staff to shift query habits from keywords to descriptive, intent-driven questions for richer search results.
Integrate Gmail and Photos with AI tools for context-aware assistance, while maintaining privacy and security due diligence.
Explore agentic features for automating everyday tasks, such as booking, research, and report generation.
Use multimodal and contextual analysis tools—like Chrome AI mode—to streamline research and knowledge workflows.
AI-powered search is rapidly evolving into a personal operating system, guiding businesses from information discovery to intelligent action with efficiency and precision. Those who leverage these tools stand to transform their decision-making processes and operational strategies in tangible, measurable ways.
Topics Covered in This Episode:
- Gemini-Powered Chrome Agentic Browsing Demo
- Chrome On-Page Summaries Feature Explained
- Google Apps Integration in Chrome Sidebar
- Cross Tab Intelligence in Gemini Chrome
- Nano Banana Image Editing Direct in Browser
- Personal Intelligence Preview in Chrome Sidebar
- Gemini Modes: Auto, Fast, Thinking, Pro
- Real-World Multistep Task Automation in Chrome
Episode Transcript
Jordan Wilson [00:00:15]:
If you listen to the show at all, you know, I'm a big advocate for take all my data, and I want my experience to be better. And I think now with some updates to Google search, we're finally seeing that. And for me, I love it. It saves so much time. Right? When your search just knows who you are, what you want. You know, I think back to how maybe I was using the web ten or fifteen years ago. And you're anything like me, maybe you launched five, ten different Google searches, and then you had so many different tabs open. And you're like, man, maybe if if only the search just knew who I was and knew what I wanted.
Jordan Wilson [00:00:54]:
Well, now that's here with some new updates to Google search, and that's exactly what we're gonna be talking about on today's show on how AI search might just be the new personal operating system in the next just interface for all of the information. Alright. I hope you're excited for today's show. I am as well. Welcome to Everyday AI. My name is Jordan Wilson. If you're new here, Everyday AI, it's for you. It's a daily livestream podcast and free daily newsletter helping everyday business leaders like you and me keep up with all the AI updates because there's a ton and just how we can make sense of it to grow our companies and our careers.
Jordan Wilson [00:01:28]:
So if that's what you're trying to do, it starts here with the unedited, unscripted livestream podcast. But to take it to the next level, make sure you go to our website at youreverydayai.com. Gonna be recapping the highlights from today's conversation and a whole lot more. So, luckily, we have a great guest, today, so you don't have to hear me, rant on search. So, livestream audience, if you could, help me. Welcome to the show. We have Robbie Stein, the VP of product at Google Search. Robbie, thank you so much for joining the show.
Robby Stein [00:01:57]:
Hey. Thanks for having me.
Jordan Wilson [00:01:59]:
Alright. Your your your title kinda speaks to itself in terms of what you do. But for our audience that may not know, you know, tell us a little bit about what your role is at Google because just search in general seems like it's changing a lot.
Robby Stein [00:02:10]:
Yeah. I mean, I spent I spent my time building new products and experiences for search. So, the whole thing, Google, you go to and, type in a question or use a camera for to ask a question as well and get information about anything on your mind now. And we're spending a lot of time, particularly the last, two years, on how AI can help expand what's possible with search. So you can really ask anything and get effortless information, whether you're bumping up against something on the street, take a picture of it, go, what is that? To you're in the car and you're having a live conversation with search to, you know, just typing something out in natural language now where you could put in, like, multiple sentences right into Google and get helpful info. So that's really how I've been spending my time.
Jordan Wilson [00:02:49]:
So there's been so many updates. You know, I do this every day, and it's even hard for me to keep up with. But, you know, maybe just bring our audience up to speed if they've been, you know, sleeping under a a Google search rock or maybe they they haven't noticed. But just give us a quick update. I mean, everything in in AI mode, some of the personalization, some new things with Gemini in Chrome. Just how much has search changed in the last few months?
Robby Stein [00:03:13]:
Yeah. I mean, it's been a pretty incredible, set of new experiences that we've been able to roll out. And, really, like I mentioned before, it's really under this vision of trying to have it make it extremely easy for you to ask anything on your mind and get information. And in order to do that, you need the best models. You have to have a frictionless easy experience to ask these kinds of questions, and you wanna be able to follow-up and have a conversation, particularly for these really complex needs where AI is awesome. You're trying to work through a problem. You want advice about something. You're figuring out what to buy.
Robby Stein [00:03:41]:
You're trying to make a decision about what you should do with something for your kids, and it really is this back and forth. And now that's all possible right in the main search experiences. And I'll I'll flag a few things, you know, we rolled out recently. I mean, one is that, Google search is now powered by the Frontier model series on Gemini three globally. So if you ask a question in Google and you see a little AI preview at the top or you go right to our AI mode, that will be powered by Gemini three models, with really best in class reasoning, understanding, and creative capabilities. So that's it's really an amazing thing to have as part of a core kind of, under the hood. And then when you actually go to go to use the product, we introduced two other things recently. One was, it called personal intelligence, which is now out in a limb in a labs experience for, subscribers in The US.
Robby Stein [00:04:31]:
And what that allows you to do is to opt into an experience where you can securely connect to your Gmail and your photos, and have these apps be context to your questions. And so you ask, hey. Where should I go, to dinner my first night on my trip? It may actually do, like, a query to find relevant information and go, oh, because you're going to New York this weekend, these are places that are in New York you're probably wondering about, and it has that context just naturally. So it's pretty incredible. I can talk about some examples later on on how we're how I've been using it, but that's pretty profoundly changed how I've used search. And the last thing we did was, you know, we're really trying to make it a lot easier to ask have a conversation with search. And what we're finding is that when we have an experience that allows for these follow ups, people report having much larger, you know, better experiences and jumps and helpfulness and people's reported feeling that, you know, they're getting the information they're looking for, and they feel good about what they're seeing. So when you on mobile now, search and get the little AI overview preview, when you go to expand it, you're now in a conversational experience where you can have a back and forth.
Robby Stein [00:05:40]:
It's really easy to ask follow-up questions, and go right to AI mode through that experience. And so that's just one of, you know, a handful of of many things that we've done. You know, you talked about Chrome, you know, and there's also new AI experiences that let you, have ask questions about what you're seeing in the browser. AI mode allows for it, and so does Gemini app. That's another big area we've been investing in. Many more to name, but there's just a handful from the last month.
Jordan Wilson [00:06:06]:
Yeah. And and maybe I'd love to just jump into the personal intelligence a little bit more in AI mode, because, you know, even kind of when I first experienced personal intelligence inside of Google Gemini, I was like, okay. Yeah. This is amazing. Not just, you know, the I think most people have come to expect now that chatbots remember your previous conversations so you don't have to repeat yourself. But, oh, when it brings in my other context from, you know, my Gmail and elsewhere, right, inside of Gemini, that's amazing. But now having that in search in AI mode, at least to me, right, all of those, you know, follow-up searches and even a lot of times, the follow-up work that we do internally. Right? Because, yeah, if you're searching something, you might have to bring up four, five, six, seven tabs, and then you're going through and you're putting it through your own personalization filters.
Jordan Wilson [00:06:55]:
And now the this kind of does it for you. Right? Like, can you change a little or can you talk a little bit just about how big of a change that can be for people who are using this feature?
Robby Stein [00:07:05]:
Yeah. I mean, I think this is, you know, a fundamental, opportunity to really ask a completely different class of questions. I'll give you an example. I connected pecan I so it's this personal intelligence project, and I connected it. And, I asked a question. I just said, what's a gym bag I should get? Because I'm shopping for a new gym bag. And in the response, it was pretty incredible. Like, using, you know, secure connection to Gmail, it was able to find other related context to question.
Robby Stein [00:07:37]:
It knew the gym that I was going to. It knew the kinds of bags I bought recently. It actually referenced a purchase I made for, like, a work bag that was kind of its brand that I really like. And so it it literally rearticulated, like, the vibe of kinds of products that I like in that category and then recommended me a bunch of brands and extensions that look they were so me. Like, it just got me. And then that would've I could've explained all of that. Like, you could theoretically type, here's who I am. Here's the designs I like.
Robby Stein [00:08:06]:
But, you know, no one's gonna spend the time doing that. And so it just allows for you to ask this kind of question, and it kinda blew me away, honestly. And there's is and and so that's one example. Another is, you know, as I was working on some present ideas for my wife, I just, like, set her name. Like, I just said, like, okay. Or or I do this with friends all the time now. I just I have a birthday coming up. I just say, like, oh, what should I get? You know, insert friend's name, whoever that might be.
Robby Stein [00:08:31]:
And they know who the per you know, the the the AI because it's connected to, you know, emails I've had with this person, even activities I've done with this person potentially in photos. It has this incredible context. And like you, I loved your intro. Like, I wanted I wanna give it as much context as possible because then the AI for me, can be uniquely helpful and actually help me with these things with as little work as possible. And, of course, it needs to be secure and privacy aware, and that's been designed from day one with that in mind. But then once that's all done, you can just let the AI just do a lot more for you, and you can ask these questions that are these kinda questions of taste open ended that you would have probably never never asked before.
Jordan Wilson [00:09:10]:
Yeah. And even when you were just talking there, it's it it seems like, are we at the point where search is jumping over into assistant territory? Right? Because, like, what you said there, that's as if you had a a a human personal assistant that follows you around and and takes notes and makes your life easier. Like, are we at that point where, you know, we can start thinking of search as being, you know, assistant or agentic? Or is there still more work to be done until we can look at search that way? I think for search, the main, focus has always been information. And so I think one area that we like to to really think about is how
Robby Stein [00:09:44]:
can we be the most helpful to you. And it's usually about when you're learning about something, trying to make a decision about something. And that's really where we play. And I think there's lots of AI products out there, and and I think many spend more more time thinking about things like productivity use cases or maybe send a million emails for me for my business. Like, search is less focused on that kinda construct. But I think over time, there's some themes that will will will cross all all AI products. You know? One is obviously in in this personalization. The two the second one is also, I think, in this category of agentic, which people throw out all the time as a buzzword.
Robby Stein [00:10:16]:
I think I think of it mostly as a way of describing the technology that can do things on your behalf as agency. So if you want to, buy something or research something, it might take you hours, in the past, but a little AI thing can go do run a process and do a bunch of research while you're sleeping and then send you a ping and say, hey. I finished your report, or, hey. I found this cool deal for you. And, again, this probably wouldn't be anything that you would personally do yourself because you're not gonna, like, scour, you know, thousands of websites looking at a specific price for every single shoe. But an AI could do that now, which is pretty neat. And so, like, in search, you know, we think about it as connected to completing the loop for these these informational questions. So when you ask about, you know, where, you know, where are cool restaurants, to eat out in New York for my trip, you don't actually just wanna know a list of cool restaurants.
Robby Stein [00:11:07]:
You probably wanna go to one. And so that's an opportunity, we think, to introduce something that we've built around, booking reservations and an agentic thing where it connects to, Talk and OpenTable and brings through live, availability in terms of restaurants in addition to the recommendations you get because we you know, the AI is able to know you well enough to recommend those things. That's pretty special, and that allows you to then complete the the the actual task you had at hand. And so those are the kinds of agentic things that we're excited about in search. Obviously, that's gonna be true very differently for other products. Like, imagine productivity tools that are designed for work. Know, they're focused very differently. Maybe they're gonna go fill out a whole spreadsheet for you, or they're gonna do a bunch of lead gen research or whatever.
Robby Stein [00:11:48]:
But we're we're really focused on this this this everyday informational use case.
Jordan Wilson [00:11:52]:
Yeah. And, you know, speaking of that and, you know, I'll ask you this through maybe two different lenses, maybe one on the the product side, you know, you you know, building this technology that billions of people are eventually gonna use and then on the consumer side. So, you you know, one thing I I I think of is, you know, traditionally using Google search, it's always keyword intent. Right? Or maybe that's just been, you know, ingrained in me for so long. Here's my, you know, my three word phrase, my five word phrase that I'm gonna put into Google. And now it seems more like intent. Right? This is what I meant or this is what I need even if I don't know. How do you prepare and build for that? I'll ask you this side first.
Jordan Wilson [00:12:31]:
How do you prepare and build for that shift kind of before it becomes needed or before people, you know, know that they're going to be searching for an intent and not just keywords? So how do you, you know, and your colleagues at Google, you you know, build that kind of infrastructure for something that really hasn't existed yet?
Robby Stein [00:12:50]:
Well, I guess, with the technology side, like you said, and then consumer side. I guess on technology side, we needed to invest in models that could handle natural language questions of of any complexity and also allow for follow-up questions, which is a different kind of technology. Because if you have a stateful follow-up, meaning it takes the context of the prior question, the models and the technology needs to be able to handle that. And that's a different paradigm than a one and done question. And so we had to invest in the technology. We had to research it. We had to do the right evals and, you know, really look at study it and say, are we able to build a system that uses all of the, you know, stuff that people want, as out of Google to develop really helpful responses, to really specific questions to your point that are not just keywords. And so that was a partnership with the Gemini, you know, modeling team, obviously, to build what became AI overviews and AI mode, and then now we have this capability.
Robby Stein [00:13:42]:
In terms of how it's implemented, I think what I think about as a product, kind of product thinker is you wanna you wanna build in the in the natural paths that already exist in the product. And so what's interesting is even late, like, kind of emerging use cases, there's always some evidence or the signature of it in the product. So for us, people were actually trying to use Google this way for a while. It wasn't a huge use case to your point. People just put a few words in typically because that's how, most people do it. And if you put in, like, a 10 sentence question into Google, you probably got a results not found, you know, a few years ago. But people would ask questions that kinda were just questions. And they could be simple ones like, you know, what's the weather tomorrow? And then it would, like, show you the one box with the weather.
Robby Stein [00:14:27]:
Right? But they did put that question in there. And so you have this base that says latent behavior as this merging behavior, I guess. And we now then showed AI overviews with those. So you now get an AI response. Then the user learns over time, oh, if I ask these natural questions that are real, like, kinda conversational questions with real English language, not keyword ease, it's gonna produce this AI thing maybe. And so people learn that, and then they do more of it. And then what we've seen is that this this behavioral this learning grows and grows and grows. And that's one of the reasons AI overviews has grown to be such a large product.
Robby Stein [00:15:00]:
Mhmm. You know, used by so many people now is because people are learning that you can ask a regular question, put that right into Google, and you'll see this thing, and then people find it very helpful. And then with AI mode, it's kinda like, oh, I want more of that, or I wanna be able to control when I get the AI. So you can go right to the AI and ask these questions. And, obviously, the next step is you people once you get a response, you kinda you have a follow-up question. Right? You wanna keep going. And now you can do that in AI mode. So if you think about it, it's like there's been meta kind of metaphors in the past around, like, a lily pond and kinda, like, are trying to build these little bridges or something like that.
Robby Stein [00:15:34]:
But there's you kinda start somewhere, and then you kinda bring someone along in your product versus trying to get them to do, like, an entirely new behavior or use a completely different product that they've never experienced
Jordan Wilson [00:15:48]:
and have no familiarity with. And, you know, it's it's funny because as you were just kind of describing that analogy, it it it kind of set off a different set of questions in even my own head. Right? Because on the consumer side, you know, even for me now and especially since some of these new personal intelligence features have been rolling out, I find myself thinking like, oh, well, should I be doing this in Gemini, or should I be doing this in AI mode? Right? And and, like, you know, from the consumer side, I guess, is there you know, do you think that there ever might be a time where someone just goes into AI mode and they're saying, hey. What should I be doing this week? Right? And just these much more open ended questions. But when it does have access, you you know, to your own kind of personal intelligence, is it ever gonna creep into the, like, oh, it's kind of could be used interchangeably with, you you know, like, the Gemini chatbot?
Robby Stein [00:16:39]:
Yeah. And I think, one exciting opportunity is that because personal intelligence works as a service in general, it works in the Gemini app. It works in search. So you should have your context with you. So to your point, regardless of what products you're using, it should be enriched with the information that's most helpful. And in Google, people typically ask informational questions. So for people, there's always billions of people using Google. If you're using Google every day, you have these questions.
Robby Stein [00:17:02]:
Now those questions can be, even more relevant to you, hopefully. And if you're using a chatbot or Gemini app, great. For people who are using those products, maybe people are using those for similar use cases or different ones, like productivity or creativity or you're running your small business, whatever that might be. That has the context too. So what's nice is you don't really need to to to to worry so much. It's kinda like use the products that you're used to for the things that you need it for, and they should be, enhanced by personal, intelligence in that context, that's now possible. So a lot
Jordan Wilson [00:17:35]:
of what we've been talking about so far is kind of the conversation and the technology and the data that goes in before you land on a web page, which is obviously a big part of search. But I do know that there's been some recent advancements to okay. What happens after you land on that web page? Can you walk, our audience through a little bit of some of the newer kind of AI powered features of, you know, that are now baked into Chrome and maybe how that might change how we can or should, use the web, when we're on Chrome.
Robby Stein [00:18:06]:
Yeah. So there's a few really cool trends happening here. One is we call it contextual search, but it's really how does the AI kinda come with you so that it can help you in the in the places you need it. And, you know, one of them is on device. The other one is in the real world. And so let's talk about on device first. So it's two two primary devices. You get your phone and your computer, basically.
Robby Stein [00:18:25]:
And on both, we we've really tried to make it easy for you to ask questions about what you're seeing. So if you're in Chrome through the, you know, the core browser kind of URL bar, search bar, you can ask a question, go right to AI mode, or, or you can get intel to ask questions about the web pages you're seeing as well. Gemini has something now where if you're a Gemini app user, and just a general user of Chrome, you can get this kind of this little kind of, drawer slides out, and you can ask question of what you're seeing as well. And so what's nice is you can now you know, if you're shopping, say, is this the best price you can find? Or, you know, if you're looking at, an article and you're confused about something, you could say, like, hey. Like, can you help explain this to me? Let's say you're learning and you have, like, a homework page up. Can you help me understand what's happening in in question three? I'm a little confused. And now, like, an AI will pop out. Hey.
Robby Stein [00:19:16]:
I can help you. And then boom. You got, like, a step by step of what to what to do, how to think about a math problem, whatever you're doing. So that's really cool. On the phone, there's incredible products from, obviously, from the Google app to Circle to Search on Android, the Gemini app that all have opportunities for you to bring your, your context right in. So if you swipe up from Android, you can ask a question about anything you see. You can circle it. A lot of time people are on social media.
Robby Stein [00:19:43]:
They'll circle something. They'll go, what's this outfit? You know? I wanna buy this this this outfit, this look. Or you someone, you know, mentioned see someone on social media as this new celebrity. Like, who is this person that everyone's mentioning? And it gives information about them. So, I think there's that's kind of a big, big trend. And then in the real world, it's through the cameras and and it's through live. And so you kinda show AI what you're up to. And so we actually have opportunities in the Google app now where you can have a live session.
Robby Stein [00:20:09]:
Like, while you're cooking, you can have your camera open, and you could ask it questions about what it's seeing, like, live, like, with you, which is pretty epic. You can also do it for homework and get help, like, right on right with you. And then you can go audio only. And, like, I do this all the time in the car, and I'll just talk to it, like, when I wanna learn about something or get smart. And I'm effectively talking to Google in the same way and having a conversation. It's all talking to similar models with similar power superpowers and capabilities, but it's kind of adapts to the the context, of your day.
Jordan Wilson [00:20:44]:
Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on GenAI. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft, and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for chat g p t training for thousands or just need help building your front end AI strategy, you can partner with us too, just like some of the biggest companies in the world do. Go to your everydayai.com/partner to get in contact with our team, or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on GenAI. And that's, you know, that's really powerful, I think.
Jordan Wilson [00:21:52]:
You know, this, like, some of those examples. So if you're listening and you're like, okay. Like, how can I start using this? I think, Rob, you just gave us, like, a great example. Right? You you essentially, you can have Google and and talk to it and go through your own, you know, layer of intelligence right there. But, you know, one thing I do have to ask you, though, because, you know, it's 2026. We can't not, you you know, talk about agents. You know? It it'd be real bad on me if I didn't ask you about this. Right? But, you know, how do you think, you know, proactive, search or, you know, even agentic search, how is that starting to, you know, reveal itself within the different Google search products, and, you know, how might we expect that to continue to change, throughout the rest of the year?
Robby Stein [00:22:35]:
For us, the way we think about it is we're trying to help you with the with usually complete the task you've you're set out to do. And for Google, the thing we see the most, people researching, let's say, restaurants, travel opportunities, booking local services, Those are the ones where when people ask questions, you wanna help person complete the loop. You know? Like, another example might be getting movie tickets or event tickets. And so our first agents have been in those spaces. And so when you ask about cool restaurants, we show you available open open table, like I mentioned availability. If you wanna look book a haircut, it can help you find availability. And if and if they don't have a website, it could even, uses really other cool AI that to call the, local business for you and then get back to you over email about what what's going on all through the agent, which is pretty epic too. So I think that's where you're seeing us invest right now.
Robby Stein [00:23:26]:
But I do think overall, if you kinda add up what that means, it means that search is becoming more capable of doing things on your behalf or doing things when you're not, you know, directly, asking it to do every single task. Whether that's research. So, you know, one question in search, gray on mode actually kicks off in the back end, like, potentially dozens of Google queries actually behind the scenes. So it's already happening where the models are doing more work for you than what you originally asked for to try to be better at solving your need, but you can kinda take it a step further and help take actions on your behalf down the road.
Jordan Wilson [00:23:59]:
Yeah. And that and that part for me, right, some of the dorks out over this stuff every day is extremely exciting. Right? And I think at least when it comes to the chatbot side, right, my own brain is now, you know, shifted to I'm orchestrating. Right? I'm I'm giving agents as much data. They're going off and doing work for me at all hours of the day, and I go back in and and check on them. Do you think we ever might get to that point on search, right, where, hey. You know, sir you you know, Google knows that you need to start looking for gifts for your wife or, you know, you need to start planning this trip. Hey, Jordan.
Jordan Wilson [00:24:33]:
This trip is a week out, and you don't have a hotel yet. Here's some things that you should be looking at. Like, is that gonna be kind of how we might be interacting with search in the future? Might it be that proactive that it's helping us solve issues before we even know that they're issues?
Robby Stein [00:24:49]:
I do think that's a huge opportunity for search to be more useful for you. I mean, I think, again, the way we think about it is it's really in the domains that people want Google to be helpful. And so for search, it's really in these kinds of spaces that I mentioned. Learning, it's education, it's, you know, asking questions about travel, food, TV entertainment questions, sports. Those are things that we see all the time, and and there's an opportunity to, let's say, you're, doing your fantasy league. Like, one idea is, like, how can we be helpful? Like, every Sunday, you gotta set your whatever. Every, you know, Friday, you gotta set your team. Okay.
Robby Stein [00:25:22]:
Well, knowing that, like, how could we have, an agent do research as an example and help make recommendations or even, like, help you just do it at some level? Like, those are the kinds of things that'll be possible soon, with the right amount of reasoning, intelligence, compute, and, of course, personal intelligence to understand you and and kinda your personal needs. And so that's pretty awesome. It's just one small example, but I think you will we have an opportunity to do a lot more for people in this age that combines kind of, like, the combination of AI, personal, but your context of, like, you and the these tasks that can do things and do research asynchronously behind the scenes. That is a pretty unique mashup of capabilities, and I and I would look for some pretty exciting products
Jordan Wilson [00:26:03]:
to come from that. Yeah. And, you know, speaking of exciting products, you know, Google's obviously been on a complete tear recently. But even some of the some of the newer updates. Right? Things like, you know, the Genie three world model and, you know, Flash being, you know, more, like, multimodal and, you know, ever just being more multimodal by default. How does this change maybe even what's possible? Right? When you do have you know, Google is much more than just a, you know, a text based, you know, or Gemini is much more than just a text based large language model. Right? How do all of these other products, you know, under the Google umbrella maybe change what's possible for search?
Robby Stein [00:26:42]:
Yeah. What's nice is because it's built on the Gemini three model series, you know, search and AI experiences in search will benefit from all that innovation. And so in the future, as the models, for example, have gotten really good at, tool use and things like instruction following. So if you ask it to do something, it kind of understands more intuitively what you want. That's awesome because that allows us to innovate much faster and be able to just through a little bit of information, have the model do more, instead of doing, like, very heavy training and other other processes that were, you know, just harder and took a long time. So that should hopefully increase the pace of innovation, and also just let it be more functional. Because a lot of these are now getting encoded more as capabilities. So let's say you can use a tool.
Robby Stein [00:27:23]:
Right? Like, so you could you could get a model and you could define a tool. You could give it an API call. You could give it a spec. Say, hey. Whenever a user asks about this question, use this API. Here's the information that's available in it, and you can return this information and great. Okay. Well, you just enhanced the whole product without changing the model.
Robby Stein [00:27:40]:
You just registered this, like, tool that does this thing for your company if you're build if you're a builder out there, or for your use case, if you're building a business or an AI service for, you know, accountants or lawyers or whoever, you could you could you could add that now arbitrarily, and you just added, like, a killer feature to your product without having to upgrade your model, fine tune your model, or do anything, which is really amazing. And so I do feel like you're gonna see this this pretty fast, this this acceleration and as as busying and dizzying as the as it is to follow AI, we'll be able to do even more than ever. But the last thing I'll say, which is pretty neat, is the user will be able to direct that more than ever as well. So if you say, like, hey. Help me learn about, lift in physics for aviation. The models now are now starting to be capable to code and actually build its own tools for you. And that's one of the neatest things with, Gemini three Pro that we've had where it's generating these little simulations, and it will code up on the fly a little, like, experience that it shows how like, a little airplane flying with arrows on lift, and it has some sliders that show, like, wind speed and all these things. And if you change them, it changes the the direction the the lift and the angle of the plane.
Robby Stein [00:28:55]:
It's, like, pretty ridiculous, like, versus trying imagine trying to explain lift to someone purely through text. Like, that's it's hard. And so the model decided it's gonna code up this little thing. And so that's gonna let users be in control and directing these technologies, and then the model will be able to to reason about and go, hey. What's the most helpful thing this user needs at this moment and just code it for you, basically? And that's a very that's a, obviously, very new frontier for us, but is is,
Jordan Wilson [00:29:25]:
oh, unlocks even more things that we don't have to design as product builders. It's kinda, like, give you the tools, and then you can you can kinda direct it as you need to. Yeah. The the, the interactive, results that I've been getting from, you know, AI mode definitely changing how I'm learning and consuming information. But, you know, Ravi, as as we wrap up here, because we've covered a lot, but maybe I wanna rewind and just, you know, as we finish here, What's your one best piece of advice for people that are, you know, looking at, you know, search, and they're like, okay. With all of these new features, I want to get more out of search. I wanna save time. What's your best piece of advice for them? I think the main thing is to get comfortable
Robby Stein [00:30:04]:
asking exactly the thing that you want with as much specificity as as you want. And I think people are limiting themselves right now. Like, if even if you're looking up, say, restaurants in Nashville for a trip, People will just write still type. They're just so used to this, kind of habit. They'll just say, you know, restaurants in Nashville, and then they wanna browse. But if you said rest restaurants in Nashville for a friend trip with six people focusing on outdoor events that I could quickly afterwards go to a music venue to, and one of us has a dog, and one of us has an allergy, and we all hate this kind of food. You can literally put all of that into Google now. And I'd and very few people probably do, but you should because now, yeah, that's where we're going in terms of the complexity of what we can handle.
Robby Stein [00:30:51]:
And I think if you wanna get the most out of Google, that's the first thing I would recommend. Then the second is there's increasingly a set of these AI tools that if you go to AI mode or in this course search box, now you can hit the plus and see kind of what's there between nano banana and these different tools like pro and different models. You can just as you have a lot of power to up you can start uploading, things. You can ask ask it to generate imagery or create an infographic for you explaining how stocks have changed or something using data that it it could find. Your creativity is your own kind of, limitation and slash opportunity, and I would encourage you to poke around just what's possible.
Jordan Wilson [00:31:31]:
I love that. Great great advice from someone helping, billions of people use the technology. So, Robbie, thank you so much much for taking time out of your day to join the Everyday AI Show. We really appreciate it.
Robby Stein [00:31:41]:
Thank you. Thanks so much for having me.
Jordan Wilson [00:31:42]:
Alright. If you miss anything, y'all, we're gonna have it all in our newsletter. So don't worry if you're walking your dog and you're like, wait. What was that Robbie said? It's all gonna be in the newsletter. So if you haven't already, please go to youreverydayai.com. Thanks for tuning in. Hope to see you back tomorrow and everyday for more everyday AI. Thanks.