Wrong on the timing, not the destination: voice search, AEO and who gets named

I got voice search wrong in 2019. Not the destination. The timing.

People did not want to talk to a speaker in the corner for real search. Weather, timers, radio. That was about it. Smart speakers sold. Ownership is not voice search.

What has changed is the brain behind the microphone. In this episode I cite Google's own August 2026 figure that 63% of Gemini users talk to it, and Gemini past a billion monthly users. Treat those as on-camera cites until we check the primary. Google Assistant is off on Android as of 4 September 2026. Hey Google now means an LLM.

We walk through the recycled "50% voice by 2020" myth, why AEO has been the house category at Question.Marketing since 2019, five guesses I label as opinions, and why every major voice still reads the same web. Publish corroborated niche truth, or none of them will say your name.

Series: Will AI Recommend You?

Full transcript: https://question.marketing/library/voice-search-wrong-on-timing-aeo-2026

YouTube: https://www.youtube.com/watch?v=JcbtIi04Sas

Transcript

Copy Transcript Copied

SPEAKER_00 0:00

So, voice search. Is it the future? I got voice search wrong back in 2019. I was telling anyone who'd listened that voice was the next big interface, especially with working with Alexa. I even co-wrote a book that had toasters in the title Speaking to Your Toaster. You know, we were thought we thought we'd have toasters listening to us. Maybe not quite yet. But people didn't really want to talk to a speaker in the corner of their room. That was the problem. You know, we had an issue with nobody really got to the point of it being the interface. They'd ask it for the weather or ask it for a timer or a bit of radio, but that was about it. But here's the comfort uncomfortable truth. I think I was wrong about the timing, not the destination. I absolutely think a hundred percent we end up talking to our AI agents or our LLMs. Sixty-three percent of people using Google, Google Gemini now talk to it directly. I myself I talk to Claude. Um, I don't talk to Siri, interestingly enough. Um, but remember, all of this time, for years, your mobile phone has been listening to you, be that through WhatsApp or through other apps. Uh Meta didn't buy uh WhatsApp and spend $19 billion on it to not have it do what it actually does in terms of listening to our conversations and building their own knowledge base. Maybe they were way ahead of the game. So that stat that I'd mentioned, about 63% of people talking to Gemini, that's Google's own stat, and it's from August 2026. And Gemini has passed over a billion monthly users now. So in this video, I'm going to talk about what happened to Voice Search the first time round and the actual numbers we're seeing today, where I think it goes over the next five years, and my guess at which AI voice wins. Somebody's going to tell me I'm wrong. Well, let's debate it. Stick it in the comments and let's talk about how voice search and AI VS or AI voice search develops over the next maybe couple of years. It might be that quick. So I'm Simon Young. Uh I am from Question Marketing. Website is question.marketing, and I'm going to talk about what happened to voice search the first time around. So if we look back around 2016 to 2019, every single marketing conference then had the same slide. Google was talking about making the entirety of Google video at that point by 2020, but 50% they were talking about 50% of all searches will be voice by 2020. And you'll still find that on stats pages today, which um tells you everything about those stats pages. But here's the nerdy bit that I love. Somebody actually traced it back. It went from an article to a blog to a campaign magazine in 2016 to ComScore. And ComScore pointed back to a 2014 Fast Company interview with Andrew NG. And he was the chief scientist at Baidu at that point. And what he said was that within five years, at least half of searches would be through images or speech, not not just voice, images or speech. A forecast from one interview got rounded off and repeated until it was carved on tablets. It was pretty much at that point what everybody thought would happen. And what did happen? Smart speakers did sell in their masses, but the marketing was probably not matching, or people bought them for the wrong reason, let's put it that way. They didn't buy them to do voice search. They became radios uh with a timer, and in the US, ownership has been stuck at roughly a third of people for years. So 33% in 2021, 35% in 2025, according to Edson research. Now fair play to us Brits here in the UK. Um the UK is ahead of the game. Edson's UK study puts smart speaker ownership at 45% of people aged 16 and over in 2025, up from 25% in 2021. So we do like a gadget over here in the UK. But owning one and searching with one are absolutely completely two different things and not the same. The older Lexa couldn't really answer anything. It read out a featured snippet or it said, sorry, I don't know that. Or you could program it to have a little bit of fun with your friends and uh get it to read out things about them. And businesses spent money on voice search SEO or voice SEO for an audience asking what time it was, right? So nobody was really asking it questions about products or businesses. So that was the hope, if you see what I mean. And I should know, I was banging on about voice too at that time. Um I was banging the drum, talking about voice search, talking about AEO, and this was in 2019, and I've got blogs and evidence to support that. And yes, I got that wrong at the time, like everybody did. We all thought everything was gonna go to voice search. So, what do the numbers say about where we are today? And I'm gonna only be using numbers from companies themselves or from people who've actually measured things because the voice stats industry is actually the Wild West, as is a lot of the AEO uh industry right now. So, where we sit with answer engine optimization, I think we thought that voice search or um those speakers, those Alexa speakers would be voice search. It's just taken longer for adaptation. So the Gemini app, as we know, has passed over a billion monthly users in August 2026, and Google says that 63% of those people uh now talk to it directly, and there's a growing group who only use voice. And for me, if we talk about I shouldn't really use the phrase the younger generation, but those that those um people that have grown up only using devices, a lot of them are turning to voice and quickly. Google switched off Google Assistant on Android phones on the 4th of September 2026, so less than a month ago from when I've I'm making this video, and he replaced it with Gemini. So every Android user who says hey Google is now talking to a large language model, not a lookup tool, right? An LLM. So think about that. All of those Android users, and I think typically Android users are probably more tech savvy than your typical Apple user, although you know I'm a PC user and I do use Apple, I have an iPhone, so um I'm I'm in one camp, right? My son, or both my sons, use um Android devices and swear by them. Apple released iOS 27 on the 14th of September 2026 uh with what it calls Siri AI, and it runs on Apple's own foundational model, and it's built with Google's Gemini technology, and it has conversation history, follow-up questions, and its own Siri app. And it's in beta um and English first, so you know, first language is English for it. So let's not get carried away with all of that. But Siri finally holds a conversation, whereas you know, previously you might have used it to say, Siri, please show me um directions to somewhere, it can now properly hold a conversation with you. But whether the take up or the usage of that has already passed, I personally think you know, the the way Siri came in and it was promoted was probably one of Apple's biggest uh drops of the ball. You know, they had the opportunity, as did Amazon, but who knows where Amazon's going because um Jeff uh Jeff Bezos uh he owns a great big chunk of the backbone of the internet, and we're probably gonna see um them making big plays. Amazon opened Alexa Plus to everyone in the US in February this year, 2026, uh, and with free prime Amazon, uh it people are having two or three times more conversations with Alexa Plus than they did with the old Alexa. So we are seeing a little bit of traction. And on Amazon's July earnings call, Andy Jazzy said more than 350 million shoppers had used Alexa for shopping in the past 12 months. So this is one of the real key points that I'm trying to make. We will move into a voice search universe as we move forward, we are all going to be speaking to devices more and more and more. Um Facebook, Muse, Meta, uh, MetaCharm, all of these things are going to make a big impact. ChatGPT reportedly hit a billion monthly users in mid-July uh 2026, and voice mode is obviously built into ChatGPT, and then we're looking at smart glasses, um, millions of metaglasses getting sold, and um that is more than triple the two million sold in 23-24, right? So we're now exponentially increasing the number of these uh glasses that are getting sold. If we're talking about seven or eight million getting sold uh in 2025, what we're finding is well, personally, I don't know about the glasses because they're now talking about the intrusion or privacy issues with them because you can be filming uh people who don't want to get filmed, etc. etc. So there's all sorts of regulation coming in around that. IDC expects about 13 million smart glasses to ship worldwide this year. Um, no screen on those, obviously, just a voice in your ear. So voice uh voice in your ear from a pair of glasses. Do we want to wear pairs of glasses? I'm a contact lens uh wearer myself. I don't like wearing my glasses. Do I want to start putting a device on my face? Um maybe the the the styles are not there either yet, so maybe it gets built into all pairs of glasses in the future, but that'd be a whole uh whole nother step. But we will be in a voice search era at that point. So somebody's gonna come at this video and say, Simon, those are all different measures: monthly users, weekly users, shoppers, units shipped. You're right, um, you can't add them up, but you don't need to. Every single one points to the same way. The assistants people talk to will all have interfaces with the LLMs, they'll all be rebuilt on these LLMs, and that's going to be within the next, I would think, 18 months. So, what is different this time? Why have we changed? Why are why am I feeling like this is the moment? You know, and this is what six, seven years past when I was talking about answer engine optimization for the first time and the fact that we were going to move into a voice search universe. You have to think those of us that work in the industry, those of us that work in marketing particularly, and those of us that use AI more than most, you know, so I will use Claude, Grok, ChatGPT, Perplexity, all of Gemini, right? I'll use all of them on a daily basis. Your average user is not doing that. So we have to look at what has changed fundamentally. Um the microphone has never been the problem because, as I say, it's always been listening to us, but it's the brain behind the data that has changed, right? The old assistants were just basically a search box that you you talked to, right? They matched your words, they talked back to you, they they could take a command or a snippet. Um if you asked them a couple of things, it would it would get confused and probably fall over, right? But the new LLMs, the models that are available now, obviously the intelligence within within them is increasing at a rate that I don't think any of us can even really understand, right, let alone comprehend at this point. So, you know, the new models they have a voice at the front end, um, and let's call it AI voice search, or as I've coined the term, A VIS, right? So because voice search carries too much 2019 baggage, that's why I've said AVIS, right? So AI voice search. Please, you know, tell me in the comments if you can, if you think of a better acronym. I'm only making this acronym up because everybody else seems to be seems to be making up an acronym as we go along, right? So I don't know where we actually end up with this entire universe. Um it's it feels to me like we are on the cusp of moving properly into voice, and that's one of the things that I want to understand from my audience. If you're listening to this and you want to understand where search is going, do you find yourself typing um more and more into your chat GPT, right? So I'm sort of seeing I'm seeing people when they start out their AI journey, they are obviously typing into Chat GPT. They're not necessarily using it on their phones as much as advanced users. Um if you're at work, are you using chat GPT a lot? Like I see an awful I think OpenAI um lead the space at the moment. Um I would you know challenge anybody probably to argue that. Um I would say probably that Grok Grok has been the product of the year, although one of the concerns is I've seen that get hacked, right, and used. So I think there will be a whole other industry coming where people are hijacking other people's AI to not just use their co uh tokens but to hijack their compute power. Um, as we move forward, we're looking at that. That will be that is definitely going to be a thing. So ask yourself: are you typing more into uh your LLM at the moment, or are you starting to use voice? Now I feel like the more advanced users and the younger generation um they're the ones that are starting to talk more to their AI. So look, you don't have to type in specific uh searches, you know, accountant Doncaster, right? You might say, I run a small engineering firm and we've outgrown our bookkeeper. Who should I talk to round here that understands manufacturing, right? Think about the difference between traditional Google search and how that phrase that I've just mentioned is completely different. Now, I probably wouldn't type it in that way because I a lot of people have been in the mindset of I can type in because that's how Google wants me. Now, Google is not the internet, Google would like to think it is the internet, but Google is the gateway to the repository repository of information which is the internet, the library of data. There are many libraries now, and each of the LLMs are starting to um gatekeep their own data, right? And this will be one of the absolutely one of the um battles that get fought over the next couple of years is who has the best data. You know, knowledge is they say knowledge is power, but uh knowledge is a currency now. That really is a thing. So these type of questions that people are going to ask with voice will be far more complex than what you would type in. So I think that's one of the key things that when we speak out loud, we will ask different questions, and you couldn't have ever done that sort of thing with a normal traditional search, and this is why I talk about the entirety of SEO has changed. It is not SEO anymore, it is answer engine optimization or generative engine optimization. I'm looking forward to the day when I can look back on this video and say, look at how all the SEO guys changed all of their titles to AEO, GEO, or they made up some other brand new acronym to go with it. Um, they're still fighting at this point. Um when we look at questions getting answered by voice, if it's on a screen, there's only really one answer, or you saw a um, you know, a traditionally a list of ten blue links that you would click on and you'd do your research, and you know, you would go through all of that and learn yourself. Um, and if it's still if it's on a screen, Chat GPT or Google AI, it will show you a few options and it will show you some sources and people talk about share of voice and citation. But if it's in your ear or through your glasses, or in the the speaker in your car, or maybe even you know a speaker on your desktop, if we end up well, the charm or the meta charm may well then speak to you, etc. Um, you might get one or two results, right? But when we go out into talking properly to an LLM, and this is why it will move into voice search, when you talk to it and have a conversation, so imagine you are having the conversation over a topic uh with the world's expert in a particular subject, you can ask anything, and that conversation will develop based on what you want, how you interact with that, call it a person because these LLMs are effectively people. So, you know, does does real intelligence then what what can you classify as intelligence? And artificial intelligence is going past human intelligence just purely because it can react quicker and it knows more, but having a proper conversation with it can very quickly get you to where you want to be rather than typing in to you know Google or to um you know into one of Chat GPT, Gemini, etc. So voice apps absolutely has to become the norm because we can get so much further more quickly, and the fan out queries that come from that and the questions that come and the discussion, so discussion will be a thing over topics. So you don't need a page two, um, there's not even a page one, and this is why I've been talking about the end of search engines, right? But then we Move into the fact that LLMs um can take action, right? So Google could never take an action for you. So you were you were then left with having to make action or take action yourself. You can take these LLMs or any of the voice agents and interact with it to the point where we will get to if I'm looking for a particular product, um, I can talk to it, get it refined. The AI agent will then become sales mode, let's imagine that. And eventually we get to a product that I want, and I just say, Can you just order that for me? And it holds my account details, it holds my bank, it pays, the product turns up. Why would we not um why would we not end up in a in a world where voice search wins? So Gemini, look, Gemini obviously can carry out actions across I think 40 or 50 uh Android apps, um, and this is you know this is according to Google. So the voice doesn't just recommend someone, it can book things, um it can then order things. Business owners should be really, really thinking about this right now, getting their businesses ready, and part of that is publishing knowledge, right? Obviously, voice is the most extreme version of an answer engine. Um an answer engine optimization is what I believe we're working towards to intelligently answer questions for users as quickly as possible. That's what Google's mantra was when it started, right? So you need to either be the answer, right, or you will be invisible as a business. So where does where does this take us in the next five years? I'll have a guess, right? So let's put a few predictions in place. And I'm gonna be fair, right? These are my opinions, these aren't facts, these are things maybe come true, maybe don't, but I'm laying this foundation down to be able to just say maybe this is where we end up. Um, and I've been wrong, I have been wrong on voice before. In 2019, I thought this was gonna move a lot more quickly, right? So you can weigh up any of these predictions um as you go along, and I'd love to hear people's opinions, right? So voice search will disappear as a phrase, not because voice dies out, because it stops being special, right? It's not nobody says touch search or type search. You will just talk to your assistant when your hands are busy and you won't type, right? You're going to talk. So maybe we should coin it as uh talk search, right? Maybe that's a thing. I don't want to start making acronyms up as we go along, right? But voice search isn't necessarily going to be a thing, it will just be the fact that you will talk to your assistant, right? Um, it's just a different brain, and it's another door into this information knowledge base that is getting built by all of the LLMs. Um, prediction two the car and the glasses, right? They will come first above your kitchen, right? You're gonna use hands-free, um, eyes busy moments, you'll be driving or walking or on site or cooking. That's where voice already wins, right? And smart glasses are the first device where where voice is the main way in, right? So we're already seeing that. Um, I would probably like to think other devices are going to challenge that, right? So people are forecasting around 27, 30 million pairs of smart glasses by 2030. I still think that that is um a drop in the ocean, right? That's obviously everybody or most people own smart devices like phones, um, and the adoption of smart glasses is slow compared to obviously we've all got a phone in our pocket. Um so it's gonna be a device you hold in your hand, or a a device you put in your pocket, uh, or a device maybe you clip on your pet lapel. Um, I don't think it'll be, I might be wrong, but I don't think it'll be the glasses, right? Um prediction three local and urgent services feel it the hardest, right? So traditionally, if we look at plumbers, dentists, etc., they've been very slow to build their local uh SEO. We're now moving into a world where this is gonna become so important, right? My prediction is that the biggest opportunity for SEO businesses, if you want to call them that, I'd rather call it AEO, GEO. Um, the biggest opportunity for people working in search optimization is gonna be working with their local businesses to make sure that they have visibility. As we stand, everybody's trying to get listed on a national level or trying to go back to the big keywords. I want to be listed for, you know, optician, just the word optician, right? No, you need to be optician in town, right? Because when people are searching locally, somebody might say, Who can you find me an optician that's open at the moment? Um, who's nearest to me? That's what's going to be important. And it's those spoken questions with one answer, and those businesses that get named uh that are going to get that work. So if you are a local business or a local trade or local service, this has to be the moment, right? This is the most important moment for you guys to grab that advantage over and above anybody else, because I can pretty much guarantee most local businesses, probably 99.9% of smaller local businesses, are not working on AEO or answer engine optimization or talk, talk interaction, or whatever you want to call that type of search. Nobody on a local level is really working on that. Prediction four voice shopping, right? This is going to work. Um old voice shopping flopped, right? Because you couldn't really buy. We already see when you're watching TV, you're watching, let's say, amp something on Amazon, you already see um the ability to interact with your tell and buy. Uh, but we are going to be moving into a world where voice shopping becomes a huge thing, right? And an agent comparing three or four quotes. Imagine seeing a product and you know, you you see it for a thousand pounds and you're like, Oh, I'm I'm sorry, um Chat GPT, that's a bit expensive, you know. Um can you just keep an eye out for when a discounts are available uh and let me know, you know, whenever you see it below £850, right? If you see that, let me know. And that's the sort of thing we're gonna be out of do. Um prediction five B2B is gonna be a slower uptake, right? Uh a finance director isn't buying a £50,000 system through their earbuds. That doesn't work. You know, you're not going to say, well, I'm looking for a piece of software that does X, Y, and Z, blah blah blah, and it you talk through it, and then it recommends you a consultancy firm and a you know a piece of software, and that's 50k. And it's just not, you're not gonna say buy that, right? You're gonna go and have a meeting and you're gonna get to know the business better. So, but the opportunity still sits there, and the research will start with a spoken question, uh, and that might be in your car, it might be to your glasses, it may be to your Muse device, it may be, well, your charm as it's known, but you may well be then talking, uh, you will be talking to a device to get you 95% of the way there in B2B sales, right? It will be slower than local. But the honest counterweight to all of this, AI referrals, let's remember, are still very small, right? And they're a very small share of website traffic, albeit that when somebody's found you on AI, an AI search, they then are five to ten times more. That's my from what I know, five or ten times more uh likely to go and convert to a customer. So in a scenario where it's normal web search, we normally see a lot of web traffic, but a small percentage uh converting. In AI, we're finding smaller volume, but a lot more of it has intent and is there like more likely to convert. We're at a point where most people are still typing, obviously we are, um, and privacy concerns are still putting people off talking to their devices. I will be publishing a video further on in this series about my concerns over privacy, um, and I think it may be one of the things that really does hold the whole space back to some extent. There's going to be maybe, you know, there's there's a lot of people that push back and they get left behind. I don't know. Um so we're not talking about act now or die, right? Anyone telling you that is selling urgency. It might sound like it's urgent. There is an opportunity, there's a huge opportunity. Um, the space we sit in is still very growing exceedingly quickly, uh, but it is a small channel, right? The window to be the answer is open, and all of the LLMs are experimenting with how they come about making you the recommended answer. It's going to be fascinating over the next six to twelve months to see how that develops. Um, and I'll be here to share how that all grows and how we experience it together. So, which of the voice with voices wins, and maybe that's why you're watching this video. I don't know, right? Who knows? Um I'll give you the runners and riders, right? And we're at you know, Gemini, Alexa Plus, Chat GPT, Siri AI, Meta AI. Gemini with its um billion users. Well, it's impressive, right? It's dangerous because of the amount of distribution it has, and it's now the default voice on Android, right? It's inside search through AI mode, and it's in the car through Android Auto, it's on the Nest speakers, and it's you know, it's also helping power Apple's new Siri, right? So Google could win twice, and half the people using it won't even know that that's what they're using. So that's one of your maybe the favorite. Amazon Alexa Plus, don't write it off. Um it owns the kitchen and it owns the shopping basket, and in the UK, it's the assistant that most people use. I don't know. We'll see. I think we're gonna maybe move into something that's in your pocket or on your lapel. Um a device sat in your corner, I don't know, right? Ofcom's figures had Alexa well ahead of Google and Siri here. Uh the question ends in, you know, do you use it to purchase? And that's where maybe I don't know, right? Chat GPT, the one people choose on purpose, right? When people want to think something through, that's often where they go. That's where we are choosing to ask the questions at the moment, right? And the voice mode is decent. Um, its weakness is that it doesn't own a phone, right? It doesn't own a speaker or a car, and it's hard hardware, is the wild card, right? And we can't tell yet. And Siri AI, finally a contender, maybe, but it's late, it's in beta, and the brains are partly Google's, right? Apple owns the iPhone and the earbuds, so I'd never count them out. But right now, Siri is a doorway, not a destination. Meta AI, Dark Horse, right? I'd probably bet that this is going to be very close to the top, one of the top three. Um giving Muse away in the US for free is a huge uh benefit. And obviously, Europe, they're looking at everybody that's got WhatsApp. So talking into WhatsApp, if WhatsApp becomes a voice AI, I think that's probably where that goes. Um Meta's got 70% of global smart glasses on the market. Um, and if glasses become the thing, I don't think it does personally. Maybe I'm completely wrong. You'll look back at this in a year and uh maybe laugh at me, but I don't know. I don't think uh it's gonna be it's gonna be um the glasses, right? So Gemini's gonna win on volume, it's the default almost anywhere. Google's Google's got the power. Can it hold it, right? It's gonna take something different and big to make a big difference. They're all gonna try, right? Um Alexa may win the shopping basket. Jeff Bezos has got his eyes on that. He cannot afford us to move into a uh voice purchase world without Alexa um and Amazon, right? It would decimate their business if shopping just went from Amazon, and they've got to have that on that's their number one board uh item every meeting, I would imagine. Chat GPT is going to win um in terms of you know the questions. So here's what I actually believe matters to you, and it doesn't really matter which ones wins, right? Does it does it really matter? Uh they all learn from the same web. The web is where the information sits, and they all go looking for the same evidence. So if you're clear uh and there's a corroborated single source of the truth in your niche, you get whichever voice is going to do the talking. And if you're not, it doesn't matter which one wins because none of them will say your name, right? You need to be listing your business now. So let's just see how it plays out. I will be maybe playing this video back in five years and having a laugh at what I've said, but um follow me if you can. Love to hear you in the comments. Let me know anything you think on the subject, and I'll see you in the next one.