AI is changing everything about apps, including something most developers don’t think enough about: how to actually make money.
In traditional software, your most engaged users were often your most valuable. With AI, they might be your least profitable. Every prompt, image generation, regeneration, and interaction costs money. And those costs can add up fast.
Recorded at Google’s first-ever Global App Summit, this episode of NEXT with John Koetsier brings together three experts to explore the rapidly changing economics of AI-powered apps:
- Ayesha Thakkar, Google product go-to-market lead for apps
- Gayatri Agrawal, founder of AI deployment company Altered
- Luu Hung, app growth and monetization specialist
Check it out:
We explore why AI is disrupting traditional freemium economics, whether advertising can actually fund AI features, how hybrid monetization works, and why subscription fatigue is pushing consumers toward new models.
We also get into real-world examples: an AI document assistant that unexpectedly became a relationship chatbot, an expense management app that uses rewarded ads to pay for AI features, and Google’s new advertising formats that could monetize the time users spend waiting for AI.
Plus, the rise of vibe coding is making app development accessible to almost anyone. What happens to professional developers when everyone can build and publish an app?
Transcript: AI is breaking the app economy
Gayatri Agrawal: I think funnily enough the most, somebody who uses your AI the most is also your least profitable user.
Luu Hung: Yeah.
John Koetsier: Ouch.
Luu Hung: Yeah.
John Koetsier: Totally. Ouch. That’s a big difference from the freemium game, Super pumped to have you guys here. We’re at Google’s first ever global App Summit, and that’s awesome. So we’re gonna be talking about apps. We’re gonna talk about monetization.
John Koetsier: We’re gonna be talking about AI, a lot of stuff about that, and so we’re gonna have some fun. You know what? It’s a big crew. Let’s start with some intros. Why don’t we start with you?
Gayatri Agrawal: Lovely. Thank you. So my name is Gayatri. I run an AI deployment company called Altered. We work with mid to large-size enterprises, and we help solve their bottlenecks, operational bottlenecks specifically, and we do that by building AI tools. I’m also an AI thought leader on socials. We help people learn about AI, you know, for free on socials.
John Koetsier: A little intro, please.
Ayesha Thakkar: So hi, I’m Aisha. I have been in Google for the past 16 years, and I am a product go-to-market lead for apps based out of Singapore, but I’ve lived across different countries in Asia. And I have the privilege and honor of working with gaming and app publishers globally and helping them to make more money using our products like Google AdMob and Google Ad Manager.
John Koetsier: Wonderful. Singapore, so you’re as much of a time zone refugee as I am right here. It’s, like, seven hours off my regular time zone, so if we fall asleep, that’s why. Your intro, please.
Luu Hung: Okay, so I’m the funny one. I’m Hung, and I’m an app growth specialist based in Vietnam. My main job is to do acquisition, engagement, and monetization for app. And besides my main expertise as an app growth specialist, I also do content myself on TikTok and LinkedIn to share about UA knowledge for free.
John Koetsier: So we’re gonna hit AI, and we’re gonna hit where AI is. We know AI is everywhere in apps right now. People are building apps with AI. I’ve done that. You’ve probably done that. We’ve all done that probably to some or lesser degree. But we’re also monetizing with AI, and we’re shipping apps with AI in them and having to monetize those, and that’s interesting. That’s a challenge. That can be quite expensive, can’t it?. Aisha, why don’t you start?
Ayesha Thakkar: Okay. So absolutely. Adding AI will of course add compute cost to your entire app development cycle, and it does change the unit economics of how you’re going to be calculating ROI and what you’re going to be measuring for. Because unlike a traditional model where you build the software once and then you just continue to monetize with every incremental user you get, with AI you could get a very engaged user, but that’s actually also going to cost you money to be able to continue to serve that user.
John Koetsier: .
Ayesha Thakkar: So AI changes the game, and I think this forces publishers to rethink how they wanna monetize and how they wanna stay profitable.
John Koetsier: So AI giveth and AI taketh away. Wonderful.
Luu Hung: Okay. I think AI help us to build app faster and also bring a lot of app feature, yeah, integrate into our app. But sometime we underestimate the price caused by AIs. Mm. When to, you want to price the cost that AI bring back to bring for you need to price for the whole journey. Mm. Take an AI, take an AI photo editor. We launch it in Vietnam. When we do it we think that the cost will be computed by the By a time.
John Koetsier: Yeah,
Luu Hung: yeah. But we see that people generate a lot of image at the same time.
John Koetsier: Oh.
Luu Hung: And they just only pay when they generate the final edition one. So that’s why, for example, user can generate five version before they … And basically you are paying for five, for five version, not for just one. So that’s why we have to change, we have to change the how we measure the CPA. It will be that it will be the CPA per the cost per generation, cost per final generation that people actually pay for it.
John Koetsier: It’s a really good insight. Yeah. It’s a great insight because we’ve come from this world of freemium economics, right? And there’s almost zero marginal cost to add a user. Almost zero, it’s never totally zero, but almost zero marginal cost. And now there is.
Gayatri Agrawal: I actually would love to double down on what he said, because I think we often forget that AI also has cost for regenerate. And usually, you know, we build a lot of products, and clients always ask us, “Okay, Gayathri, you know, what’s the operating cost like?” And usually, you know, we used to do this, which is we used to ask them, “Okay, how many documents are you parsing?” Or- … “how many people are gonna be using it?” But we often don’t realize that, you know, they plan for probably 100 documents, but they do it multiple times. Because, you know, you don’t always get the right answer.
John Koetsier: Oh, AI’s not perfect?
Gayatri Agrawal: I mean … Right? And I feel like what people really miss out on is they need to not just look at the operating cost from the pers- from the perspective of successful generation but also how many takes do you need to reach a successful generation?. I think that’s extremely important for everyone to also calculate while, you know, calculating their operating cost.
John Koetsier: I think that makes a lot of sense.
Ayesha Thakkar: Overall, compute cost is going to go down. Mm. So I think we also have to bear in mind that we are in this phase right now, but you know, we are continuously building newer models. We’re continuously finding newer efficiencies.
Gayatri Agrawal: .
Ayesha Thakkar: And there is a lot of optimization that even developers as well as, you know, even providers that Google is bringing on. So I- I feel like this is going to continue to be an evolving and dynamic space, so none of these answers are going to be, you know, sort of ground in stone forever.
John Koetsier: I- we need to underline that a couple times, boldface it, maybe put a highlighter on that because I think a lot of us when we start getting into AI and apps and development, we’re using pretty good models, and we realize pretty soon that you don’t always need to use a frontier model or even a mid-tier model for everything. You can often use a finely tuned, good, but very cheap model- Yeah … for a lot of stuff, and then you can actually make some money on it.
Gayatri Agrawal: Yeah. Which is also redoubting your task pieces, how complex they are.
John Koetsier: .
Gayatri Agrawal: So, but you know, like to add to what Aisha said, I think funnily enough the most, somebody who uses your AI the most is also your least profitable user.
Luu Hung: Yeah.
John Koetsier: Ouch.
Luu Hung: Yeah.
John Koetsier: Totally. Ouch. That’s a big difference from the freemium game, and- Yeah … that’s a big difference from the IAP game as well. You were gonna say something
Luu Hung: Yeah, because I have, I have an experience about this. For example we have a, an app and we integrate a chatbot AI into this app. Basically it is a utilit- utilities app. It help you to scan the doc, import the doc, and do something with your document. But we add an extra feature. We call it AI chatbot, and to answer for you everything related to your docs. Love it. So yeah we want to add some more value proposition to make the user acquisition better. But when we track the CPA the CPA who the CPA of which user- Mm … cost a lot and we compare it with the frequency they use the AI- … chatbot. And we see that is totally not- Not proportional … not at what we expected. The one who use AI a lot is the one who will not pay for this. The model here is hybrid. Yeah. We do with IAA and IAP, but the IAP contribute more than half of the- Mm … revenue stream.
John Koetsier: .
Luu Hung: And so that we want to make people to do with IAP. And when we check for the user who use AI a lot, AI chatbot a lot to see which job they are require AI to do for them, and surprisingly, they ask AI to chat, to have a social chat, have a casual chat. Oh, wow. Not specifically, not specifically related to the document they are uploading. Maybe we are doing the wrong way because we use a model of AI to integrate in this- Yes … and basically it can, it can answer you everything- Mm … not only help you to process with the document. Oh. It is like we spoil the user. Not spoil the user, but user, they are doing it in not the right way that we are expecting them to use. And it cost up a lot of money for the, for the token we have to pay.
John Koetsier: Yeah. It’s super interesting because you had a utility, and it became a relationship app. It was an AI boyfriend or girlfriend or partner or whatever, right? That’s hilarious. How did you solve it? Did you create a harness for your AI? Did you, did you create something for like, s- some guardrails for your AI to only ask que- answer questions in certain areas?
Luu Hung: Actually, we had to build a local one a local AI one, and just limit the question, the answer, and the q- the question and the answer- Yep Within our knowledge not using the third party anymore.
John Koetsier: Mm.
Luu Hung: And it’s much more cheaper because at the first time we underestimate the cost bringing by the chatbot.
John Koetsier: You did market research, and you found out there was a valid and massive market for things other than were in your actual app And you could create another app for that. . Interesting. I like it.. Okay, so we’ve gone to subscriptions, right? Though how many subscriptions can you support? Like, how many little cuts and tears in your wallet do you want that are leaking money all the time?
Gayatri Agrawal: I mean, honestly, and I say this as an AI creator and founder, I have a lot of subscriptions, and I’m a pro user of- What’s the
John Koetsier: number? What’s the number? Give us the number …
Gayatri Agrawal: think more than 20.
John Koetsier: Holy mother. Have you, have you asked Muse to cut a few of those?
Gayatri Agrawal: Yes. My job is to test everything out, right? But having said that I think there’s definitely subscription fatigue. There’s a lot of people getting worried about if they’ll be able to catch up, if they’ll be able to learn. We also train, let’s say, leaders and other people on AI.
John Koetsier: .
Gayatri Agrawal: I’ll tell you the question that all of them ask, right? Which is, “Look, I’ve learned this tool today, but what if I have to get another subscription tomorrow, and then maybe another day after, right?” I think the world is going towards aggregators also, which is You know, you see like lots of models, lots of use cases on one app, I think which is also why tools like-
John Koetsier: Super apps.
Gayatri Agrawal: Super apps. But I’m a big believer of, like, having, A, the ecosystem. I feel like for everyone, when they’re already in the Google ecosystem using things like Google Studio Flow and things like that, just puts everything together. It can access my Drive and Docs… So I think ecosystem matters, and second, aggregators because I don’t think people for the longest time will pay for multiple apps together. Not everybody is like me. They
John Koetsier: pay- You’re single-handedly supporting the app ecosystem. Yeah. Thank you so much. Developers of the world are very grateful. Yeah.
Ayesha Thakkar: Sorry, I was gonna just jump in and say that’s also a great opportunity for publishers and developers to think about how do they make the most of the folks who are staying on their app. Mm. Because not everybody’s gonna subscribe, not everybody’s gonna pay, but you want that scale because that’s also how you stay in business, and- Yes … you stay in your top grossing or top downloads charts. So I think bringing in hybrid monetization as a way- Yeah … of being profitable is a great way of being able to monetize the user base who’s there but not choosing to subscribe or pay. . But you’re able to actually make money by showing them ads at relevant you know, interfaces or as part of the user journey. And actually, we have so many different types of ad formats that you could actually find the right moments to monetize in that journey. And to your point earlier, where you were talking about we may not really need a very heavy tier or a mid tier model, I think it’s very interesting if you even do the math for how can you use ads or rewarded ads as a way to balance out the cost. And I think with the lightweight models, it’ll actually even pan out. With the heavier models, you might want to have more of a hybrid model where you do get subscribers. But with the lightweight models, ads can be a very sustainable way for you to monetize and continue to scale.
John Koetsier: Can we double-click on that in a second? I wanna come right back to that, but I wanna … we had 20 subscriptions right here. How many do you have?
Ayesha Thakkar: I actually don’t know
John Koetsier: Oh, no. It is a lot. You haven’t asked AI to figure it out for you?
Ayesha Thakkar: I haven’t asked AI to
John Koetsier: figure it out
Ayesha Thakkar: for me. It’s like
John Koetsier: $3,000 flying out of your account every month?
Ayesha Thakkar: No, it’s not that bad. But I would say I think per category, I probably have two or three. Per category. So it depends on, like- Wow … entertainment and, Slayer. … billing and yeah, I mean, you know, as a consumer- Okay. Okay … there are a lot of interesting contents out there- Yes … that you wanna be able to- Yes …
John Koetsier: observe. The app ecosystem thanks you as well. How many subscriptions for you?
Luu Hung: Actually, just a few. I think three, four.
John Koetsier: Oh, wow.
Luu Hung: Yeah, I can name them.
John Koetsier: Wow.
Gayatri Agrawal: It’s really optimized.
Luu Hung: Because I do not want to be, I want to be in control-
John Koetsier: Yes …
Luu Hung: for what I’m paying for.
John Koetsier: .
Luu Hung: Because I don’t want to be kind of surprised for unexpected billing from subscription. Yeah, I experienced myself in the past. So that’s why I closely monitor what I subscribe.
John Koetsier: Good
Luu Hung: for you. And I think this is the behavior that very common in Vietnam.
John Koetsier: Yeah.
Luu Hung: Now-
John Koetsier: I think it’s common elsewhere as well.
Luu Hung: Yeah.
John Koetsier: But I wanted to double-click on what you’re talking about, right? So there is subscription fatigue. Not everybody wants to have 20 subscriptions, but we wanna hit a wider audience, and there’s multiple reasons for that. You- you- you- you mentioned a couple of them. First of all, if you have this very narrow niche app and you only have a few people, it can be profitable for you because it’s high value, but you’re never gonna rank in any of the app stores or Google Play or anything like that because you just don’t have the volume that’s gonna take you to the top there, and that limits your future growth, right? So that’s one reason. Another reason is, which we’ve had in freemium economies forever, is we don’t wanna just monetize 1% of our users or 3% of our users. That’s just wasteful, right? Talk about what you are starting to get into on hybrid monetization and … ‘Cause that’s a new thing for me. I didn’t think that you could do admon or hybrid monetization for AI ’cause I just thought the costs were too high Go a little deeper, if you would.
Ayesha Thakkar: So, let me give an example. So we are seeing different verticals of apps who are already incorporating this. So I don’t know, are you familiar with drama shorts, short drama?
John Koetsier: I am familiar. I personally don’t.
Gayatri Agrawal: I’m sure you must have
John Koetsier: seen She wants to look at my phone. She’s gonna check on me.
Gayatri Agrawal: No, because I’m sure everyone’s, like, attached to those billionaire wives and husbands, right? So-
Ayesha Thakkar: I’ve heard about the billionaire wife who was- I mean, they’re very addictive stories. So to give an example, right? Yes. Like, I think they are the ones that I feel have done such a good job of incorporating different monetization models depending on the different type of content that they have, depending on the different type of market that they are serving. ‘Cause you also have to keep in mind user behavior and disposable income changes respective, you know, where you go… And so what I’ve loved about seeing their growth is they started out being a lot more IAP and subscription first. Yeah. But as they’ve scaled out globally, as they’ve kind of expanded their content development platforms and they’re bringing in new content, you know, every week, every day they want to, at that scale, be profitable and be sustainable.. And they don’t want to only monetize, like, 5% of their user base, even in mature developed markets, right? Yeah. So with bringing in ads, they can actually control … For example some of them typically will say “Watch an ad to unlock an episode.”
Gayatri Agrawal: .
Ayesha Thakkar: And that way they are balancing out the cost. And then just to take a bigger picture perspective, it’s not just about monetizing that, but they’re also thinking about, as a developer, how are they optimizing their entire content production flow.
Gayatri Agrawal: Mm.
Ayesha Thakkar: So how are they actually producing the dramas that they’re showing you? And they’re using a lot of AI in that as well- Mm … to cut costs. Mm. So if you take a sort of bigger picture view, you know, you can actually incorporate AI not just in the monetization part or the production part, but even in your developer operations, even in your monetization operations. And then you have a far more sustainable ecosystem that allows you to scale globally and touch users giving them the flexibility of choice in terms of what kind of experiences do they want to choose or pay for.
John Koetsier: Yeah. That makes a ton of sense, and I think we saw on the first day of the conference, the Sensor Tower presentation, that there’s two massive major markets which happen to be two of the top 10 growing markets in India and Pakistan, that it, ad mon is the primary mode of monetization. And it’s only a few other markets, the US is one of them, and there are some other nations on there as well, that are IAP mostly or subscription, right? So if you wanna grow- Then you’ve gotta have different options. Any thoughts there on
Luu Hung: that? Yeah, I really want to double down, double down on the idea about give the users the flexibility of option. Because taking an example of Vietnam we have we have some kind of like expense management app, and for the expense management app we devise a revenue stream is very clear that we will we will live by IAP model. Because for the IAP model is more sustainable because user not only pay for this w- not only pay one time, but they will pay to unlock the feature, and the revenue, the lifetime value is much more better. But otherwise, we need the IIA model. And for the IMA- IIA model, we design the journey like that. For the early adapter, early adapter who we, who are our target audience are Gen Z, and they are start to learn how to manage their expense.. So you can– they can come to the app, and we integrate an AI feature that it is an OCR feature of the AI. So basically, it means that you can scan your bill scan your transfer statement in your bank account and then the app will automa- automatically detect and input all the expense- Nice … that you have. And that is a feature we sell for the mass user out there. And the objective of this is to– We have two objective. The first one is to educate the user, but we do not want to waste our money on educate the user a lot. They- … because every time they use the OCR feature, we have to pay a small amount- Yes … of token for this. So that’s why we initiate with the IIA model, so they can watch the reward ad to unlock one more OCR-
Gayatri Agrawal: Mm
Luu Hung: chance to scan for the bill. And we define that if they unlock if they use the OCR feature for four times a day- Mm … so they are very close to the IAP. So whenever they use the OCR more than four times a day, we will pop up with the ask to see if they want to subscribe to use unlimited OCR.
John Koetsier: Yes.
Luu Hung: We use the word unlimited OCR, but we measure that. Basically, user only use five to six time a day for the OCR, so basically we are still profitable if user do the IAP. And we balance this with the revenue from the IIA, and then we try to convert the IIA user, the, Sure … free user to the paid user by tracking their behavior.
John Koetsier: Is that a loss leader type of model? Like, are you getting as much out of the in-app ads as you would’ve out of the in-app purchase? Or is it a little bit less and you’re kinda saying, “Hey, over time they’ll, they’ll graduate”? Yeah. How does it work for you?
Luu Hung: It is like, when they use the freemium feature for a period of time-
Luu Hung: they will get used to with it, and they will try … They are starting to try to input more expense-
John Koetsier: Ah …
Luu Hung: to the app. For the, at the very first time they just include like, they just input like one or more one or two of their costs, they expand into the app. But- Mm … regularly they are trying to input more, and more, and more. And for the IAP feature, we have the forecasting for the wealth forecasting feature.
Ayesha Thakkar: .
Luu Hung: So that we calculate the income and the expense of them, and we calculate how they save each month, and we can forecast for the, for the certain amount of saving each month, so how long they can achieve some target like housing, car- Yeah or a trip. And yeah- It’s- They love that feature.
John Koetsier: Very sophisticated. That’s very cool. Is any of that apply to your world?
Gayatri Agrawal: So more than me running ads, I’m a consumer of ads being run on me.
John Koetsier: .
Gayatri Agrawal: I think two things you’re right. One we’ve been seeing, and all of us know, that India is a very difficult market for subscriptions or buying apps. India does not pay on a monthly level as much as compared to other countries. I think one thing or one thing, like one feature I’ve really, like, seen work is when people actually go for in-app payments via Google Pay- Mm … ’cause India is super big on quick payments.
John Koetsier: Yes.
Gayatri Agrawal: So you know how Apple Pay is, like, great for outside of India in international markets? India is super big on Google Pay.
Luu Hung: .
Gayatri Agrawal: So if you get them to pay via credit cards and debit cards, you see a lot of friction, and you see a lot of, like, people, you know, backing off. But I have seen some of the biggest, you know, AI companies figure this out. Instead of adding, let’s say, a $20 subscription which equals to, let’s say, approximately 2,000 rupees, they make it the $20 equivalent of Indian rupee, Which is not literal calculation, right? .. And then they make it much lower, they put it on Google Pay, and I think that’s how you’re being able to, like, get people to buy apps and subscriptions. Yeah. That’s definitely been one of my, like, biggest observations. So for everyone watching, if they’re trying to, like, launch an app in India, I think it’ll be important to just check out the buying patterns, you know, in our country, where people are able to pay very quickly via Google Pay. That’s one. And second, I think one interesting, like, in-app purchase that I’ve been seeing is a lot of these apps, right, be food delivery apps and quick commerce apps- Mm … they’ve actually started building cohorts of consumers.. And they see that, okay, if a tier one consumer is buying from us, they sort of calculate, right, that, hey, you know what? They probably want to get educated, and maybe selling them an AI course will really help. So they partner up with companies- Yes … who sell AI courses or, like, who sell, let’s say, beauty creams and things like that, depending on the cohort, and then they run you know, those ads- Mm … just when you buy the product.
John Koetsier: . That’s getting bigger and bigger in North America as well, and has been for a year or something like that, right? You’ve got a massive cohort there, and they might wanna buy other things, and you can expand your share of wallet is basically what’s going on, right? I wanted to get back to the economics of using ads to fund AI features in your app. Is it hard? Is it possible? You mentioned different units. You mentioned different ways of doing it. Thoughts on that?
Ayesha Thakkar: Maybe just to break it down, I’m sure the actual computation of this is a lot more complex than what I’m gonna be able to explain now. But I think fundamentally when you incorporate ads into your monetization model, the way to evaluate whether this is working for you or not is to look at what are your earnings through these ads, right? And each of these impressions will have a different value associated based on the different type of format, based on the different type of market based on the different type of session length.. So I think it has to be taken and calculated together. So like, Hung was mentioning, you also need to look at the retention aspect of it, right? Because is this user coming frequently enough that you are able to monetize them multiple times with, say, different types of ad units? Then the actual revenue that you are earning from that user is actually going to be quite different from a user who’s just gonna come in once, and then you’re not seeing them for the next 30 days. So I think the way to kind of look at it from an unit economics point of view would be to think about what are your costs for running the AI query or the AI model that you’re using, and what kind of optimization can you do for that, on one hand. The other is to look at what are your earnings at scale, at impact on retention across the different sessions. . And then I think fundamentally sort of underpinning all of this is what is the value proposition? Are you genuinely providing something, you know, uniquely useful that will make them want to keep coming back? But I mean, this is, this is the type of calculation that even we at Google have to make. Yes. You know, we just announced a bunch of AI features at the App Summit in how we’re bringing AI in our AdMob interface with Ask AdMob, how we are bringing AI in our coding environments via the AdMob MCP server. But we are also thinking about what is the value we are bringing.. Because we know that if you use the MCP server, you’re going to have to use your tokens-
Gayatri Agrawal: Yeah
Ayesha Thakkar: To actually, you know, pay for it and do all of the queries. When you use Ask AdMob in the AdMob interface, we are going to pay for the tokens there. Yeah. So we wanna make sure that- … the value is real- Yeah … for the publisher, but also for us as a provider of these tools to be able to be sustainable and scalable at the scale at which, you know, we operate and we make these tools available.
John Koetsier: How many quadrillion tokens did we see on stage that Google’s using per month right now? Was it– It was some ungodly number. Wow. I wish I had that unlimited token budget, don’t you? I think we’ve covered a ton here, and it’s been super interesting, super amazing. I wanna give each of you sort of like a last crack at maybe something that I didn’t ask, maybe something you had burning in your heart that needs to be released, and let’s have a chance to talk about. We’ll start with you, and we’ll go down.
Gayatri Agrawal: I think you know, because we build a lot of models for different organizations the idea is to also be able to provide them the value. And I wanna double do- double down on what Aisha said, which is around MCP. I feel like we’ve been looking to work with a lot of different products, get the data, and then build on top of it. So I think anybody out here building AI models, being able to give access to, you know, developers like us I think that’s really interesting for me because we are being able to work with organizations’ data. We’re then being able to get all the frontier models’ you know, access, and then build something together on top of it. That’s been a very exciting journey for me. I think to add on that, because we’ve spoken so much about, like, value and monetization and even subscription, I think when we work with different enterprises, the same question that’s asked by all of them is, “Hey, but what’s the value proposition that AI brings it brings into this process of ours?” And the way we’ve been able to do that is by saying that, “Hey, today, you know, this is your cost of operation for this, you know, part of the process. This is the operating cost as discussed earlier of AI, and you know, this is the value or ROI we’re being able to provide in terms of time and money.”
John Koetsier: Makes sense. Absolutely. Every business has to think about that. Aisha, anything from you?
Ayesha Thakkar: I’d say two things as well. The first I would say is I know we spoke about hybrid monetization, but I’d like to double click on just a couple of points with respect to the types of ad formats. So I think we’re very used to seeing ads as a banner ad or an interstitial ad that pops in between levels or between, you know, actions that you’ve completed. But I’m, you know, very excited to see what publishers do with some of the new type of ad formats that we’ve designed, which are more user-friendly, which are non-disruptive as long as you implement them correctly.
John Koetsier: Isn’t an ad supposed to be disruptive?
Ayesha Thakkar: No, you can actually make them flow actually far more naturally into your user journey. Good. So for example, some of the new ones that we launched called Picture-in-Picture and Squeeze Back, they are partial screen ads.. So these are formats that the publisher can configure when they show up, and the user is able to dismiss them entirely so they can go back to doing whatever they were doing. Or you’re monetizing a moment in time when the user is idle.. You’ve entered a prompt, you’re just waiting.. And you’re able to kind of, see an ad in that moment in time- It’s the lock screen of
John Koetsier: AI. …
Ayesha Thakkar: and then go back. And these are the innovations that we need to think about, right? Because if you want to be able to monetize, but you also don’t want to get dinged on reviews or on being annoyed you wanna think about, okay, how do we evolve these ad experiences? And we definitely want to make sure that the ecosystem benefits which is the publisher makes the money, the user gets the content, and you know, the utility of the app, and providers like us are able to build that platform for them to make that happen.
John Koetsier: Makes sense. And
Gayatri Agrawal: can I, can I add
John Koetsier: something? Wow. Another one? Really? No, you’re allowed.
Gayatri Agrawal: Yeah, no, I just wanna add to this because, you know, so many people are making memes on the internet about how, you know, they press enter, and their AI is, like, working. They’re chilling in the park. So but I think it’s a very important thing to notice for all the people who wanna run ads, like you mentioned because it takes time for the AI to sometimes build you an app or a feature, and I think that time is extremely important, and maybe you can make use of it and run ads there. And in fact, like, there are- super like, you know, funny extensions available, like Google Chrome extensions available in the market which allow you to like get paid if you were to look at the ad while the prompt is, you know, generating your answer. Okay. But maybe, you know, it’s a, it’s a thing that we start seeing- Maybe you could watch a
Ayesha Thakkar: short drama, I think.
Gayatri Agrawal: Yeah. Interesting though.
John Koetsier: Very cool. Usually when I’m waiting from one thing from an AI, I’m getting another AI to do something else, but last hits, last topic.
Luu Hung: We have already talked a lot about the benefits that AI brought to the publisher and also for, to the to the user. But I think I want to frame the role of AI in another angle because I see this is a trend in Vietnam right now. AI create a lot of opportunity, a lot of benefit, but also create a lot of competition. It creating- Mm … a lot of yeah, competition landscape in Vietnam. Yes. For example, we have so many, so many people who are zero in technical skill, and they are flooding and do Vi coding. And-
John Koetsier: The number of submissions to the app stores- Yeah … has been increasing vastly.
Luu Hung: And un- unexpectedly they are doing it very well. Yeah. We have a group in Vietnam some group in Vietnam, it is some group called like, Vi coding in Vietnam, build in public, something like that. And there are so many showcase of the user who are the Vi coder. They said that they have zero technical knowledge, and they are building an app, and they are monetize from the app. And some– And I see some case studies they can monetize an app just 100 … I remember 50,000 to 100,000 in just one month.
John Koetsier: Wow.
Luu Hung: And that is the first time they do this. And, That bring me an that bring me a talk that I think this is now the new era, the new era of publisher. Not the publisher previously, if you want to build an app, you need to have developer.
Ayesha Thakkar: Yeah.
Luu Hung: And sometime you are very confident about your app features, your app value proposition. It is a high barrier, high ent- entry barrier for other publisher to entry because they need time to code for an app. But now the speed of release an app is m- is very quick. Yeah. And even a people with zero knowledge can release an app. And if they have the user experience message, because I want to double down on the part of I- IASA. If you w- if you have the user experience mindset and you can use the ad placement, you fit in into the user journey, and then you combine it with the features that you provide along the way. For example, if people– if y- you know that people really want something at the end of their journey-
Luu Hung: so you place an ad or you place an, a paywall before this, you know, they are gonna pay for it or they are gonna watch your ad anyway.. But if they have the good mindset about user experience and also on the ROI, return on investment, you– they know that for feature cost you need and the return they get, so I think they will be sometime, they will be something that very challenging for the publisher out there.
John Koetsier: Yeah. I think that’s a great place to end our little chat because it has– AI has democratized who can build that. Yeah. But that interesting part for me is when with that reality of everybody being able to create and build, which is a wonderful thing in a sense, the professional publishers have to do what Ayesha was talking about a little earlier, have to double down on quality and user experience, and just make this excellent in every way possible so that it stands above the crowd. It’s fundamentally easy to be mediocre today, right? Almost anybody can be mediocre. Mm. And to be excellent- Yeah … is still challenging. I wanna thank you guys for taking this time. It’s been a wonderful conversation. I really appreciate it.
Gayatri Agrawal: Thank you. Thank you.
Luu Hung: Thank you so much.
John Koetsier: Thank you.