Where Does AI Go Next? An Interview With Android Boss, Sameer Samat

We may earn compensation from reviewed products. Learn about our editorial policies.

I don’t often get to sit down with the people who actually build the platforms we use every day. But last week, I had 20 minutes with Sameer Samat – President of the Android ecosystem at Google – and I wasn’t going to waste it on safe questions.

AI fatigue (the big one for me, and I know a lot of you out there, too), agentic phones, Gemini in the car (what will it do differently?), cross-device wizardry that actually makes sense to normal people… we got into all of it, and Sameer was extremely forthcoming with his answers.

Below is the interview in full, lightly edited for this website. You can watch the full interview on the Mark Ellis Tech Diary YouTube channel.

AI fatigue – is Google listening?

Mark: “For the last few months I’ve had the impression that there’s AI fatigue setting in. People appreciate what it can do for productivity, but a lot of it is just being thrown at them. What is Google doing to make that tech fall into the background and give people their time back – while still delivering the actual benefits?”

Sameer: “I think you’re bringing up a great point. Look, as a technologist, what computers can do now with AI is amazing – it’s incredible, it’s exciting. But that’s because I’m an industry nerd. Ordinary consumers have never really wanted to focus so much on the technology. They want to focus on: how does this help me with my day-to-day life? That’s the key. And if it helps me with my day-to-day life, that’s wonderful – I’ll use it.

“If you want to talk to them about the technology, there’s a time and a place for that, but usually that’s not every day, all the time. So I think what we’re trying to do is focus on the features and capabilities that can really help you lead a better life. Google has always tried to be a product that gets you to the answers to the questions you have as quickly as possible – and I think YouTube is helping people do that now too.

“That’s what people really appreciate on Android. Some of the things we’ve tried to do are bring this technology in a very purposeful and intentful way, where we can solve a problem that people have every day. Let me give you one small example. We implemented a feature a couple of years back called Circle to Search. It’s not Circle to AI – it’s Circle to Search, because that is something people really understand. They’re looking at the purse their favourite influencer is carrying, or a beautiful picture of a beach somewhere and they can’t describe those things as images to Google to learn more about them – because that’s hard. They can now long-press the home button, circle it, and that initiates a visual search. It’s an amazing feature. There’s a ton of AI going on behind making that happen, but we’ve intentionally focused on: how does it help you? How does it enhance your day-to-day life? That is the approach we’re taking with all the features we’re launching with Gemini and everything else on Android.”

Wondering how AI is shaping up on Google’s hardware? Check out our Google Pixel 10 review for more.

---

The barbecue test – agentic AI in practice

Mark: “In a recent LinkedIn post you wrote about an example of AI in action during a BBQ – very much up my street, especially with the weather in the UK right now. Can you walk us through that – from speaking to your phone, to what it actually does, to what you get out the other end?”

Sameer: “I’ll give you the specific example and then back up a little bit and explain how it’s a good reference for how we’re trying to change the way the phone works in general.

“The capability I was speaking of is called app automation. Instead of doing all the taps and swipes yourself for a complicated task, you can simply speak or type what you want done, and the system will go figure it out – actually launch the app and go through the steps you want.

“So: I was holding a barbecue. I had a list of the people coming, their dietary preferences, how many people in their family, how many kids. We’ve all built that list. What I needed to do next was build a grocery list. So I asked it to put everything I’d need for the barbecue – according to the number of people and their dietary preferences – into my cart in the grocery delivery app I use. I just asked it to do that. It opened Instacart, looked at my list, found the local grocery store, and started adding everything.

“You can watch it do this – you have full control, you can interject at any time. Or you can minimise it into a window in the status bar so it’s happening in the background while you do something else with your phone. It’ll come back to you if it needs help – it’ll prompt you: “I’m not sure about this, can you answer this question?” It goes all the way to the end, adds everything to your cart. Then it brings you back. You get to look at everything, make sure it’s right, and you hit checkout. It won’t complete the transaction for you.

“That saves me seven, eight, nine minutes – which isn’t a lot in the grand scheme of things, but honestly, that’s seven, eight, nine minutes that made my life easier. I can go do something else.

“So if I take a step back: what are we really aiming for here? Android is an operating system, but we think we need to move from the world of operating system to what I call an intelligent system. Historically, for the last 40 years in computing, it’s been about you as a micromanager of your computer. You literally think what you want to do and break it down into micro tasks – I’m going to tap this, swipe this, move the mouse here. You’re in command of every single little piece. We even go to class to learn how to use the machine.

“An intelligent system instead takes your intention – your goal – and says: “Okay, I understand. You want to plan this barbecue. Let me go orchestrate the things needed to make that happen.” It’s conceptually a simple idea, but it’s actually a profound change in the way computers work. That is the transition we’re trying to lead from an Android perspective.”

---

Staying in control – how Google is building trust

Mark: “I’m a kid of the ’80s and ’90s, so I’ve always had that sense of control over what I’m doing on my computer or phone. This is a big change. How is Google designing this so the user still feels in control?”

Sameer: “That’s the key, isn’t it – how do you introduce these new capabilities in a way where the consumer feels it’s a positive thing, where they feel trust and safety? That’s the hard part, beyond all the underlying technology.

“We’re starting with steps that are really important. We’ve limited this capability to a certain set of apps – it won’t do anything with your banking app, for example. And even inside your e-commerce or grocery app, it’s not going to complete the transaction for you right now. It brings you back to that screen and you complete it. There’s a set of things we’re doing that are designed to make sure you’re in control.

“One of the other examples of literally being in control is that you can watch it complete the task for you. You may not want to – it defeats the purpose in some ways – but the point is you always can. It’s never doing anything in a hidden way behind the scenes. You can always watch it, and you can always press a button to say, “I’m going to take over from here.”

“We’ve also been rolling this out slowly. We started with the Galaxy S25 earlier in the year, we’ve gotten additional feedback from consumers, and we’ll expand it in the upcoming weeks and months. Listening to user feedback and taking it from there – that’s really important.

“One other thing I wanted to mention: we previewed a feature at Google I/O and at the Android Show called Rambler. Rambler is, in some ways, a reinvention of the keyboard – and it encapsulates this idea of an intelligent system really well.

“The keyboard is a perfect example of what I was talking about with micromanagement. You think of what you want to say, break it down into words, then break those words down into individual keys. That’s just how we’ve all worked with keyboards. But a different way to conceptualise it – if the computer is capable enough – is to open the microphone, and instead of it being a dictation microphone where every single word is faithfully transcribed including your ums and ahs, you just say what you’re trying to convey and it formulates it quickly into the paragraph or sentence you’re actually trying to get to.

“So you could say: “Hey, go to the grocery store, and when you’re there, if you don’t mind, could you pick up some fruit – maybe some bananas, apples, oranges. Actually wait, never mind, we’ve got plenty of bananas, don’t worry about that.” And instead of faithfully dictating all of that, the bananas piece is just omitted. That’s a different form of intelligent system – it just makes things so much simpler to convey what you’re trying to get across.”

---

Screen time, self-control, and pause points

Mark: “What fascinates me – and it must be quite a challenge for Google – is how you keep people engaged with their phone in this age of agentic AI. The phone is doing more for you; you can put it down, tell it to do something and walk away. So you’re using it less. How do you strike that balance between hands-off tech while still encouraging people to pick up their phone?”

Sameer: “I think about tech from the standpoint of: how do you make sure your use of it is not only intentional, but that you feel really good about it afterwards?

“I’ll give you an example. Even before phones, you’d sit on the couch and binge-watch a show – one episode after another. If at the end of it you feel great, like, “That was exactly what I wanted to do today, I had so much fun” – thumbs up, good for you. But if at the end of it you feel like you wasted time, you didn’t do the right thing – we’ve all had that feeling of using something too much and just not feeling good about it.

“From an Android perspective, I really want us to build software and experiences that help the user achieve that first feeling, and give them tools to avoid the second one.

“One new feature we’re introducing is called Pause Points. I’m really quite passionate about this. We introduced a digital wellbeing set of features years ago – the idea that you could get a dashboard showing your weekly screen time and set limits for yourself. That was helpful. But we’ve learned a lot since then.

“Pause Points goes beyond setting limits that you can easily turn off and defeat yourself. You can set, on a given app, something like: “When I open this app, I want to solve a crossword puzzle before I can use it” – or “show me ten pictures of my dog before I open the app.” These little moments of friction that you impose on yourself – they’re completely self-created – but we see that people want better controls over these things. They want ways of coaching themselves. We want to provide those.

“And there is a setting in Pause Points where, in order to turn the feature off, you actually have to reboot the phone. Again, it’s something you choose – but when we talk to consumers, they want some of those things. They want to know they have the tools. We think this is really important to help people with the balance.”

Curious how Samsung is approaching AI on its own devices? We put the Samsung Galaxy Watch 8 series through its paces.

---

Gemini in the car

Mark: “Let’s talk about cars – I know we share a passion there, having first met at Silverstone a few weeks ago. Gemini is coming to the car, and when it comes to in-car tech, I’m quite old-fashioned – I get in, I put music on, and I drive. Whenever new tech is introduced into the car, I really have to be convinced. What are we going to see from Gemini in the car that’ll make people think, “I could have done with this ten years ago”?”

Sameer: “I’m very excited about Gemini in the car for a number of reasons. Most of us have experienced using some kind of assistant in the car – “call Mum” or “navigate to this place.” And those assistants of the past have taught us how to speak to them. You have to use precise phrasing. We’ve all had the experience of repeating yourself over and over with slight variations because it’s just not getting it.

“With Gemini, it understands natural language. You can just speak. You could say: “Hey, I’m looking to go to the airport. I’m not entirely sure which terminal – I think it’s Terminal 2, but it’s the one American Airlines flies out of.” And it will figure that out and navigate you there. That’s very helpful, especially in a hands-free environment where I don’t want to be looking at the screen and inputting.

“One thing I experienced recently that was quite delightful: I was in a new city – Cincinnati – in a rental car with Android Auto projected, which was great. As I was driving through the downtown, which was beautiful but somewhere I didn’t know much about, I asked: “Can you tell me what all these buildings are?” I didn’t really know what it was going to do. But it basically started narrating a tour for me. It said, “Up on your right there’s this building, and up on your left there’s this one, and here’s the history…” I said, “What’s the stadium over there? Who plays there? Is it the soccer team, the football team?” Pretty cool.

“The final thing: I also listen to music in the car, but I have quite a long commute, so I do listen to podcasts. One feature I’m really excited about us building – we have a product called Notebook LM, which can turn any set of documents into a podcast. I’ve always wanted that to be ready and available on the fly. So if I was commuting somewhere and wanted Gemini to build me an on-the-fly podcast about a particular topic I wanted to learn about – we’re working on that for the car. I think that’ll be really exciting.”

---

Cross-device – the moment it clicks for normal people

Mark: “So we’ve got cars, phones, watches, eyewear – so much cool, interconnected stuff that I’m very lucky to get to play with in my job. My friends and family who don’t work in this world broadly understand the concept of context moving across devices, but we’ve gone far beyond that, as you know. What’s the first cross-device moment that’s going to genuinely wow a regular person?”

Sameer: “One of the things I’m really excited about is – as you’ve probably heard – we’re introducing a new category of laptops later this year called Google Books. These are Android-based laptops that are a fusion of Chrome OS, the Android UI, and Android together. We’re really excited about that combination.

“The laptop and the phone are a particular pair. Who hasn’t had the experience of: “I took this picture on my phone and I want it on my laptop” – and then you’re emailing it to yourself, or fiddling around with Quick Share, or AirDrop if you’re on a different ecosystem? We’re fusing the file systems together in a very seamless way, so it will literally look like you have one file system between both devices. When you go on the Google Book to open a picture, all your photos are just there – from your phone. And vice versa.

“I’m actually using a Google Book right now, so I experienced that just last night – and every time I do, I’m just like… okay. It’s such a simple thing, but it’s kind of awesome.”

If unified file systems across devices sounds appealing, see how we got on with the M4 MacBook Air in our full review.

I’d like to say a big thanks to Sameer for his time, and to Google UK for organising this interview. It was genuinely one of the most open and thoughtful conversations I’ve had with someone in his position. I hope you found it useful, too.

Let me know in the comments which part of this resonated most with you – and if there’s a question you wish I’d asked him, drop it below and I’ll bear it in mind for next time.