Artificial intelligence isn’t just about knowledge or creativity anymore; it’s about getting things done. When I say, “I need to be in Lisbon by Friday,” I don’t want a list of available flights or things to do in the city. I want an agent that checks my commitments, finds the most efficient route, reserves tickets, syncs them to my calendar, arranges airport transfers, and notifies my team of my travel schedule automatically.

That’s intent turned into outcome — and that’s the heart of the new agentic AI wave.

Agent 101: What it truly means

Forget the buzzwords for a moment. Agentic AI isn’t another chatbot or automation script. It’s AI that takes initiative: perceiving the goal, planning how to achieve it, and acting within the real world context.

As a CEO or any other senior business leader, imagine having two deputies.

The first one is an expert. They’re knowledgeable and creative. They can pull up some great stats, create a thorough report, turn it into a nice presentation with beautiful charts, and give you a list of people to send it to. But ask them to do this and more themselves, and they procrastinate for a month or get stuck somewhere along the way.

The second type is a doer. You just mention an idea to them, and they’re already pushing it into action: pulling up the report from the expert, arranging a meeting with key stakeholders, agreeing on the next steps, and then executing them. The next thing you know, you’re signing off a launch plan — exactly how it looked in your head.

This second type is a perfect embodiment of agentic AI. Except, of course, that agentic AI isn’t a person.

This shift matters because it changes how people and technology collaborate. Instead of reactively feeding or creating data, agentic systems work toward a goal: adapting, reasoning, and acting when the unexpected happens.

The good: Where technology truly delivers

The promise of agentic AI is simple but profound: it frees people from digital friction and mundane actions. Some companies are already getting this right.

For instance, IBM’s AskHR automates more than 80 routine HR requests, freeing HR staff to focus on improving employee experience and tackling higher-value, creative challenges.

As the company markets it, “imagine being able to automate the entire pipeline of approving time-off requests for multiple employees with just a few clicks. Or automating the entire onboarding journey — from collecting paperwork to setting up accounts — without lifting a finger.”

In 2023, Lemonade showcased the power of agentic AI when its autonomous claims bot set a new industry record by settling an insurance claim in just two seconds. The system instantly verified the customer’s policy, ran fraud detection checks, and approved the payout — all without human involvement.

Earlier this year, CrowdStrike unveiled Charlotte AI Detection Triage, marking a major leap in agentic AI-driven cybersecurity. Operating with customer-defined, bounded autonomy, Charlotte AI automatically triages security detections with over 98% accuracy, eliminating an average of 40 hours of manual analysis per week.

These examples aren’t just demos or prototypes. They’re live systems already proving that, when built with the right intent and scenarios in mind, AI can take real responsibility.

The bad: When AI implementation derails

For every beautiful success story, there’s a cautionary tale. Agentic AI or any other type of technology, misapplied, can magnify trouble just as easily as it removes friction. And sometimes, that’s even before they can live up to the promises of true ‘agentic AI’.

Last year, Air Canada’s AI chatbot gave wrong refund advice; the court ruled the airline legally responsible. While this was merely a chatbot, not a full-fledged AI agent, it’s easy to imagine the impact multiplied if it started booking the wrong flights at scale.

In one of the most notorious agentic AI blunders, Zillow launched a bold “iBuying” initiative through its Zillow Offers division, aiming to buy and resell homes via algorithm-driven purchases, but ended up suffering losses of over $300 million as its models struggled to predict real-estate market fluctuations and renovation bottlenecks. The venture was ultimately shut down, serving as a cautionary tale of how automation and data-driven systems can falter when confronted with complex, real-world domains.

A Replit AI user reported that the company’s AI-powered coding assistant went rogue, deleting their production database, creating around 4,000 fake users, and even lying about its actions. Tech entrepreneur Jason M. Lemkin said the AI ignored repeated warnings not to interfere with production code during vibe coding sessions.

Replit’s CEO, Amjad Masad, called the incident “unacceptable” and confirmed new safety features like automatic separation between development and production databases to prevent similar issues. The case highlights growing risks with AI coding assistants, which can accelerate development but still pose reliability and security concerns when used unsupervised in live environments.

All these failures have something in common: agency without accountability. In short, you can’t outsource ethics, empathy, or complex judgment to code.

The real: Where it’s actually working — and needed

Now, here’s a catch: Gartner predicts that 40% of all agentic AI projects will be sacked by 2027. And the reason is they’re either too expensive or not really needed. I personally believe the actual number will be much higher.

Every new technology goes through two major phases. First, rapid development and an avalanche of various implementations. Second, a kind of ‘innovation hangover’ in which only the fittest survive.

That’s what happened with smart home tech when it first emerged, for instance, soon followed by a myriad of weird devices with pointless connectivity features. Sometimes all you need is to push a button on a kettle.

Beyond the hype, real agentic AI needs to find the middle ground: quietly solving everyday problems where it actually helps. This would lead to real wins where AI saves hours and makes the difference, not headlines.

Here’s my take on a few simple rules that make building innovations around agentic AI practical and failure-proof:

1. Start with the job, not the model. You don’t “implement AI”; you solve a task.
2. Design for trust. Every agent should disclose what it’s doing — and why.
3. Localize. One-size-fits-all agents fail across cultures.
4. Test small. Build one agent for one task. Learn, refine, repeat.
5. Measure outcomes, not hype. Track time saved, satisfaction gained, or risk reduced.

Or, as I often say: Stop running in every direction. Start by finding the real job to be done.

The good, the bad, and the real: The rise of agentic AI was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

How many times, as a marketer, have you found yourself wanting to jump on that one trend, just to find yourself buried underneath layers of approval or creative delays? And by the time everything was ready, another trend came knocking? Yeah, me too.

This frustration is common in travel marketing, where complicated and expensive advertising campaigns often have to contend with erratic and unpredictable travel behavior.

Traditional marketing strategies are too slow for today’s hospitality and travel businesses to keep up with all the trends. But there’s an alternative.

What is agile marketing

Agile marketing is an adaptable marketing strategy that prioritizes allocating resources toward high-value projects, which are determined based on changing trends and consumer needs.

Stemming from the mid to late 1990s software development projects, agile approaches occurred as an alternative to traditional waterfall management methods. The latter relies on determining rigid goals and ways of achieving them from the start, without the possibility of altering and moving between different strategic stages. Agile strategies, in contrast, offer PMs more flexibility and the prospect of rapid iteration.

In 2012, noting the potential of agile approaches in marketing, a group of 35 industry specialists from San Francisco gathered to modify the Agile Manifesto for Software Developers for marketing purposes.

The result was an Agile Marketing Manifesto, outlining 7 values and 10 principles that help keep up with the complexity of the modern marketing world.

The main ideas of the Manifesto emphasize focus on customer value and business outcomes, identifying and delivering them early, learning through experimentation, fostering functional collaboration, and, most importantly, encouraging change over static plans.

Benefits of agile marketing

Since their inception, agile strategies have proved to work quite well.

According to the Standish Group’s Chaos Report (2020), which compared the success rates of traditional versus agile methodologies from 2013 to 2020, agile ones win by a high margin, with a success rate of 42% compared to just 13% for traditional (waterfall) projects.

When it came to marketing, the advantages of agile showed in:

1. Flexibility reflected in the iterative and short work cycles, known as sprints. Sprints make it easy to react to shifting project priorities and respond to sudden customer needs. Consequently, marketers can adjust the messaging, platforms, or strategies and positively impact the campaign performance.

2. Efficiency, which came from teams clearly allocating tasks and areas of responsibility. That helps marketing teams focus on the most critical work and deliver more value without needing extra effort. Like in the Pareto principle, where 80% of the value comes from just 20% of the work.

3. Collaboration, facilitating cross-functional relationships and efficient communication, reducing the risk of departmental silos.

4. Most importantly, higher customer satisfaction. One of the most important principles agile marketing stands for is the prioritization of customer needs. To satisfy them, marketers resort to more targeted and relevant programs, creating the ‘right things’ vs ‘more things’.

Why agile marketing matters for travel brands

The travel field is inherently unpredictable. More than other industries, it is vulnerable to natural disasters, climate shifts, political tensions, and pandemics, among many other factors. Because of these forces, demand in the sector is neither stable nor linear.

Just this summer, for instance, the industry experienced a rise of close-in bookings, where travelers decide on a destination only a few days before departure.

Travel Weekly reported in August 2025 that booking windows had shortened, as their reader survey found that 36% of respondents were experiencing a higher-than-usual rate of last-minute bookings. Given short booking windows, marketing strategies built on long-term assumptions won’t work the same way as agile ones.

What’s more, considering the rapid transformation society experiences virtually in every industry due to the AI integrations, fully Agile teams are proven 3x more likely to be successful adopting the new tech, as per Agile Sherpas’ 2025 State of Agile Marketing Report.

Applying agile marketing in the tourism industry with AdTech

So far, all the information and statistics provided in this article have prepared marketers to accept agility as a new norm, mainly on an emotional level. But modern advertising technology makes it far easier to both adopt and implement this agile mentality in practice.

Here are some of the instruments marketers now have at hand:

1. Geo-targeting to alternative destinations

One of the 2025 travel trends, according to Expedia, is detour destinations, which all experienced an increase in searches over the past year.

63% of consumers say they are likely to visit a detour destination on their next trip.

Travelers may come to Paris, but detour to Reims. They may come to Dubai, but detour to Abu Dhabi. The reasons vary from avoiding crowds to flight changes to just seeking something exotic. Travel marketers must be prepared to meet those travelers when they first begin to look for a detour and suggest the best options through geo-targeting.

2. Ride and mobility data integration

Ride-hail platforms, aggregated mobility networks, and location analytics give marketers near real-time insight into where people go (airports, transit hubs, attractions). Some destination marketers and airlines already embed ride-phase ad units.

Yango’s ecosystem, for instance, could connect ride data (or location signals) to ad delivery (e.g., showing ads to users en route), as Yango already combines map, app, and ad layers.

3. Real-time creative swaps

The backbone enabling many of these tactics is real-time creative swaps. Instead of static banner sets, an innovative engine ingests price, route, inventory, and context triggers to serve the message that best resonates in that moment. This keeps campaigns fresh, relevant, and predictive rather than reactive.

4. Mobility-based retargeting

Even when the trip, or the supper, or the shopping spree ends, the engagement with the audience shouldn’t. Someone who checked and visited Yango Maps for “attractions near Sharm El-Sheik” can later be shown a “48-hour resort offer” ad across the Yango ecosystem. These mobility-linked audiences consistently outperform static demographic segments in terms of visit lift and conversion rates.

Final thoughts

From fixed campaign calendars to adaptive real-time marketing, travel advertising has come a long way. Yet the industry now faces its biggest test, as traveler behavior is erratic, booking windows are shrinking, and third-party data is no longer reliable enough to predict what comes next. Traditional methods can’t keep up with that pace or uncertainty, leaving hospitality brands reacting after the moment has already passed.

The answer lies in marketing built for movement. Agile strategies, powered by AdTech platforms like Yango Ads, allow hospitality and travel brands to read signals as they happen and adjust instantly, whether that means shifting spend to active markets, swapping creative in real time, or retargeting travelers on the move. In a sector defined by unpredictability, agility is the only sustainable way to stay relevant; the industry must accept this as the norm.

Agile Marketing for Hospitality and Travel Businesses was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

A few weeks ago, I came across a discussion on Reddit where a developer wrote something that immediately resonated with me:

“It’s tough to monetize when users are in and out quickly — but adding small incentives or gamification helped extend sessions and open the door for upsells.”

If you’ve ever worked on a utility app, you’ll understand this pain point instantly. Unlike social or gaming apps, where people spend minutes (sometimes hours) scrolling or playing, utility apps often serve a single purpose. You open a VPN, clear your storage, scan a document… and within 10 seconds, you’re done.

That’s great for users — it means the app is doing its job. But for developers and monetization teams, it creates a challenge: how do you build sustainable revenue when user sessions are so short?

The Nature of Utility Apps

Utility apps are the “quiet workhorses” of mobile life. They’re essential, reliable, and often invisible until you need them. But because sessions are quick, traditional monetization strategies like long ad breaks or aggressive paywalls rarely fit.

This is why the Reddit comment struck me — it reminded me that even in quick-use apps, there are ways to create engagement moments that feel natural, not forced.

Strategies That I’ve Seen Work in Utility Apps

1. Micro-Value, Micro-Transactions

Users may resist committing to a $9.99/month subscription just to scan a few documents. But they’re more open to smaller, contextual offers:

• Watch a 15-second ad for 3 more scans.
• Pay a $0.99 “day pass” for unlimited VPN.
• Unlock a one-time premium boost for faster performance.

The secret is aligning monetization with the immediate task, rather than asking for commitment upfront.

2. Light Gamification

Gamification doesn’t have to turn a utility app into a game. It’s about nudging users to stick around just a little longer:

• A streak tracker for productivity tools.
• Rewards for consistent VPN usage.
• Progress dashboards that celebrate milestones.

These small touches extend sessions and create natural points to introduce premium features — without breaking the “get in, get out” flow users love.

One of my favourite samples for this can be the app Fortune City: Expense Tracker by SPARKFUL INC. (Note: this app can be considered as both utility + fintech)

Fortune City turns your budgeting into a city-building game — assigning buildings to your spending categories as you log transactions

Also, I was recently recommended a book: “Gamify Your Tasks: Kick Life’s Ass” by David Joel Stephens. It’s a refreshing take on how to transform everyday responsibilities into engaging, rewarding experiences. While it’s written for personal productivity, I think many of its principles could be adapted to utility apps. Imagine turning routine actions — like backing up files or completing a health check-in — into rewarding milestones.

If you’ve read it, I’d love to hear your thoughts on how this framework might apply in the app space.

3. Smarter Upsells at the Right Time

The worst moment to ask someone to upgrade is the first second they open your app. The best? Right after they’ve hit a limit or just completed a task.

AI is helping apps predict these high-intent moments — for example, surfacing an offer when a user runs out of free credits, instead of showing random popups. The result is higher conversions and happier users.

4. Hybrid Monetization Models

The most successful utility apps I see today aren’t betting everything on a single monetization model — they’re combining multiple approaches:

• Ads that are lightweight and optional (rewarded or native placements).
• Micro-subscriptions for short-term or feature-specific access.
• Premium tiers for heavy users who want the full experience.

This flexibility ensures that every type of user — whether they only open the app for 30 seconds a week or rely on it daily — finds a monetization path that feels fair. It also stabilizes revenue streams in markets where ad performance or subscription conversion might fluctuate.

The Takeaway

Utility apps may never have the long, immersive sessions of gaming or social platforms — and that’s okay. The goal isn’t to stretch usage artificially, but to design monetization moments that feel like a fair exchange of value.

That Reddit developer was right: even something as small as a reward or gamified milestone can turn a quick 5-second visit into a chance for deeper engagement — and yes, sustainable revenue.

As builders and business leaders, we should keep asking:

• Where are the natural breaks or friction points in my app?
• How can I turn those into opportunities for users to unlock more value?

Because when monetization feels like part of the experience — not an interruption — it stops being a challenge, and starts becoming a growth driver.

Monetizing utility apps when users come and go in seconds was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

Over the past few years, the AdTech industry has been gearing up for a cookieless future. And while technically, cookies are still around, the shift is already well underway: brands are moving toward first-party data and direct user engagement.

In emerging markets, this transition opens up a unique opportunity. Infrastructure here has followed a different path from the very beginning, shaped by localized ecosystems and platform-native user behavior. This now creates a fertile ground for a new kind of advertising — one that’s built on real intent signals and owned data, rather than anything else.

Cookie-based targeting is losing ground. Data from ecosystems wins

Back in 2019, Google shocked the market with its plan to deprecate third-party cookies in Chrome. And despite recently being paused, the influence of cookies is clearly fading even without a ban. Safari, Firefox, Brave, and DuckDuckGo have been blocking third‑party cookies by default for years. GDPR-driven consent requirements cut cookie usage on key publisher sites. What’s more, according to Usercentrics, 26% of users have disabled third-party cookies in their browsers, reflecting growing privacy concerns.

Not only on the web, the shift has taken place on mobile, too, especially on iOS, where privacy updates like App Tracking Transparency and the rollout of Link Tracking Protection in iOS 17 have made user-level tracking increasingly opaque.

According to Singular’s Q2 2025 Trends Report, global mobile ad spend jumped 40.3% year-over-year — but the growth hides deeper tensions.

Acquisition costs continue to rise, with some Android verticals like Shopping and Fintech seeing steep CPI increases. Meanwhile, iOS remains dominant in revenue, especially in gaming and health & fitness, despite lower install volumes.

And with ATT opt-in rates hovering between 1.4% and 9%, especially in sensitive categories, marketers are increasingly operating without user-level signals.

What we’re witnessing isn’t a sudden cliff. It’s a gradual erosion of traditional tracking methods. Attribution is less precise. Retargeting is losing its edge. So brands have been learning not to rely on third-party signals alone and shuffling for new performance signals.

Where can they find them? Well, emerging markets can give great insights.

MoMo, Zalo, Grab, and Shopee

In Southeast Asia, user behavior is often shaped by super-apps — gigantic platforms with high-frequency services like maps, digital wallets, and local media ecosystems, bundled into a few familiar names. In my home country, Vietnam, most people have these four super-apps installed on their phones: MoMo, Zalo, Grab, and Shopee.

MoMo — Vietnam’s leading fintech super‑app boasts over 31 million users and more than 200 integrated services, from money transfers, bill payments, QR-code in-store purchases, ticket booking (cinema, flights, trains), and donations, to financial products like insurance, micro‑loans, stock trading, and loyalty programs.

Zalo — Starting as Vietnam’s most popular messaging platform (~78 M monthly users), Zalo now includes audio & video calls, social statuses, file sharing, reminders, and ZaloPay integration for payments. Users also access mini‑apps like traffic info, public services, and digital wallet features.

Shopee — Southeast Asia’s dominant e‑commerce platform with in‑app browsing and secure checkout via its solution — ShopeePay, flash deals, promotional vouchers, and integrated logistics, delivered mainly by motorbike.

Grab — Originally a ride‑hailing and food‑delivery service, Grab has built out its ecosystem to include transportation (motorbike, car), food delivery, logistics (courier), and digital payments via MoMo partnership, edging closer to true super‑app status.

This is very different from the Western world, which relies on standalone services like mentioned Google Pay, Netflix, Spotify, or Uber. And though the super-apps infrastructure has some built-in fragility, it also has an immense advantage — the amount of direct, first-party behavioral data.

Any brand advertising there gets qualified users through real, observable signals. Instead of bidding blindly into aggregated segments, brands can tap into intent-rich patterns: where people go, how often they move, what they listen to, and what content keeps them coming back.

Cookieless future is when relevance beats reach

To explain how this new way of cookieless advertising works, I’ll take our own ecosystem by Yango Group as an example. It offers a wide range of digital services, including ride-hailing, delivery, e-commerce, mapping, video and music streaming, and more, with one login and one wallet, allowing users to access all services with a single account and a unified payment system. All are powered by AI and real-time data.

So, imagine a user frequently takes Yango Taxi to and from the airport within a short interval — say, two to three days. They’re likely a business traveler. This pattern enables travel brands to serve offers for insurance, hotel bookings, airport transfers, or foreign exchange.

A user who regularly orders groceries through Yango Delivery on weekday evenings likely shops for daily essentials. This data can inform FMCG or CPG advertisers to serve timely promotions, new product launches, or loyalty offers based on actual consumption rhythm.

A user frequently browses restaurants, pharmacies, or gyms on Yango Maps — and sometimes even navigates to them — indicating intent to visit. This behavior helps local businesses or big chains deliver personalized coupons, new location announcements, or targeted seasonal deals based on real location interest.

Users who spend time in Yango Play’s mini-games often engage with interactive content and micro-rewards. Brands can integrate native promotions into gameplay or offer exclusive in-game bonuses in exchange for trying out services, from free delivery vouchers to ride discounts, turning playtime into high-intent conversion moments.

The list of examples is endless. But the main thing here is that advertising on ecosystem infrastructure with deep insights into user behavior helps brands build meaningful connections based on:

• first-party data from real behavior, not assumed interests,
• localized formats designed for how people commute or spend,
• category depth in verticals like gaming, travel, and finance,
• platform-level insights that reflect what people do, not just where they click.

In regions where digital life runs through ecosystems, what matters most is not how many people you reach, but how relevant you are when you do.

And relevance builds trust.

Final thoughts

The future of digital advertising isn’t just about smarter targeting — it’s about smarter data ownership.

Targeting is no longer centered around users as digital fingerprints. It is built around context, behavior, and real needs — all surfaced through owned platforms.

When an ecosystem controls the full journey — not just the ads, but the services — it doesn’t need to borrow signals. It can build relevance natively and deliver value without violating privacy.

Contextual insights. First-party behavior. Data is processed within the platform, not passed outside it. That’s not just compliant. That’s a new model of performance — one that earns attention, not just buys it.

In markets like Vietnam and Southeast Asia, success will go to platforms that:

• know their users through real interactions,
• offer local relevance,
• and build on trust, not tracking.

That’s what Yango is doing — and why this shift isn’t just about compliance. It’s about building something better.

The future of advertising is not only about technology — it’s about who owns the relationship with the user. In emerging markets, that relationship is built through services. That’s where platforms like Yango can lead.

From cookies to super-apps: How emerging markets are redefining ad targeting was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

Along with my peers and colleagues, I always monitor the latest developments in business and technology to see what’s coming up, how the industry is changing, and what dominates the headlines. I sometimes challenge myself to look up something new and exciting that’s not about AI, and it’s getting more and more difficult. If you live in 2025, you can’t deny it: AI is everywhere.

In thinking about how the AI world itself is evolving, I see that there are certain structures that shape it, intersect with it, but also compete with each other in terms of influence and practical implications.

Let’s talk about the three pillars of the AI world now: competition to build the most advanced model, scaling up the massive infrastructure needed to run and train models, and finally, driving AI adoption by building truly useful, practical AI products.

Brace for the model race

The main competition in the AI field is now happening between the so-called foundation models — massive neural networks trained on gigantic datasets to handle a wide variety of tasks, including generation of text, images, code, and video.

There are more than a hundred models in use, but the top ones are well-known: OpenAI’s GPT-4.5, Anthropic’s Claude 4, Google’s Gemini, DeepSeek’s R1, Meta’s Llama 4, Alibaba’s Qwen, xAI’s Grok, and Stability AI’s Stable Diffusion for images.

These models are the “foundation” for a wide array of apps based on AI — chatbots, image generators, smart assistants, and various utilities. Their adaptability, scale, efficiency, training methods, and ability to work with different types of data (for example, text, images, audio, etc.) set them apart and power much of today’s AI innovation.

There’s fierce rivalry among the tech companies behind these models, and the scope and speed are astonishing. According to the AI Index Report 2025, model scale continues to grow rapidly — training compute doubles every five months, datasets every eight, and power use annually.

At the same time, performance gaps are closing: the Elo system (a method for calculating the relative skill levels of players that has been adopted to evaluation of AI models) shows that the skill score difference between the top and 10th-ranked models fell from 11.9% to 5.4% in a year, and the top two are now separated by just 0.7%.

It’s too early to say whether the foundation models have exhausted their potential for scaling and significant improvement, but there’s certainly less of a margin between different models now. And, starting from the end of last year, there are some distinct voices, including Bill Gates, saying that AI development overall may have reached its peak. In this context, some advocate for smaller, more specialized models that are well-equipped to handle particular tasks.

There are multiple reasons why this could happen, but some of the most frequently cited are lack of available data, computing power, and energy. For example, a single query to ChatGPT consumes about 10 times as much energy as a single Google search, according to the Electric Power Research Institute (although this claim was put to a test at Epoch AI).

Foundation needs infrastructure

The big established and emerging players behind the foundation models are well-known companies, modern superstars of the AI technology scene. However, their breakthroughs wouldn’t be possible without the companies that build and provide the infrastructure for AI development at scale.

Building a foundation model like GPT or Llama requires a massive investment in infrastructure. According to an estimate from Epoch AI, the cost of training frontier AI models has grown two to three times per year for the past eight years. The researchers also expect that training of the largest models will cost over a billion dollars by 2027.

First, you need clusters of powerful hardware, typically hundreds or thousands of high-end GPUs or TPUs, supported by fast networking (such as 100 Gbps Infiniband) and large, high-speed SSD storage to handle massive datasets and frequent model checkpoints.

Data infrastructure is essential as well: you must be able to collect, clean, and efficiently load huge (often petabyte-scale) datasets, and you need systems that ensure data quality, privacy, and security.

Third, for model training at this scale, distributed machine learning frameworks (such as PyTorch with DeepSpeed or Horovod) are needed, together with orchestration tools like Slurm or Kubernetes to manage compute jobs across many machines.

Additionally, specialized monitoring systems are needed to track system health, resource usage, and training progress, while regular backups ensure that work is not lost in case of hardware failures.

Finally, substantial physical infrastructure is required to support such large-scale operations, including reliable power supplies and advanced cooling systems. All of this must be managed by a skilled team of machine learning engineers, data engineers, and MLOps specialists.

In the US and many other parts of the world, cloud platforms like AWS, Google Cloud, and Microsoft Azure are the go-to providers of infrastructure for foundation model training.

NVIDIA is dominant in hardware (GPUs and networking). Their dramatic growth to a trillion-dollar company was almost entirely driven by the extensive (and expensive) use of chips needed for AI training.

In China, companies like Alibaba, Baidu, Tencent, and Huawei also provide the full stack for AI infrastructure, supplying computing power, storage, and software platforms needed for training foundation models, and they continue to invest heavily in domestic chip and cloud technology development.

Besides that, they also develop proprietary AI platforms. For instance, this March, Baidu announced its first AI reasoning model to compete with DeepSeek, another Chinese company that shook the AI industry earlier this year.

Ironically, one of the main reasons why DeepSeek’s model was such a big disruptor was that it could achieve results similar to ChatGPT’s while consuming much less infrastructure and computing power.

All this is why the companies behind the foundation models are seeking ways to increase their models’ efficiency, while also reducing the costs and dependency on infrastructure. In the meantime, though, infrastructure providers still benefit from higher costs of AI development.

Focusing on practicality and agentic AI

There’s growing interest and optimism among users about the benefits of AI. Data from the AI Index Report 2025 suggests that, in countries like China (83%), Indonesia (80%), and Thailand (77%), strong majorities see AI products and services as more beneficial than harmful. In contrast, optimism remains lower in Canada (40%), the United States (39%), and the Netherlands (36%). However, sentiment is shifting: since 2022, optimism has grown substantially in several skeptical countries, including Germany (+10%), France (+10%), Canada (+8%), Great Britain (+8%), and the United States (+4%).

AI is already being widely used across the board, whether for entertainment or productivity. From creating beautiful AI avatars and listening to AI-generated music, to writing personalized performance reviews with ChatGPT and letting Zoom’s AI Companion help you through conference calls, AI has become extremely helpful at assisting us with basic things.

As big tech is struggling with scale, pressure for innovation, intense competition, and growing infrastructure needs, we’ll likely see the focus shifting towards more practical implementation of technology in its latest evolution: agentic AI.

OpenAI and their agentic AI solution Operator exemplify this shift toward a more practical, product-driven, and empowering approach — unveiling AI’s ability to go beyond just searching, analyzing, and generating data to perform real-life actions.

Soon, AI will go beyond just giving a good answer or doing a few basic tasks. It will be able to act using reasoning and logic. AI agents will use contextual knowledge about you, the entire set of data on the internet, and reasoning to put it together and convert it into a meaningful action. You’ll be able to say something like “Book a table in one of my favorite restaurants” or “Find and order a gift for my daughter’s birthday” and get it done.

Therapy and companionship, personal and office productivity, education, software development, creativity, marketing, entertainment, smart assistants, healthcare, public services — these are just a few areas where I expect to see more exciting agentic AI products in the next couple of years.

The current AI landscape and how it’s shifting from models to agents was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

Advertising has long evolved at the plea of the masses. From print to interactive digital formats, it has found ways to adapt to new audiences, their behaviors, and medium preferences.

But the new generation of media consumers has brought the industry to a final reckoning with everything it once understood about promotion.

In fact, advertising, as traditionally conceptualized, might already be dead.

What’s blooming, though, is anything that doesn’t look or feel like an ad, yet works precisely because of it.

The youth dictates

Like many revolutionary movements, the demise of advertising is happening under the discerning preferences of modern youth.

Gen Z, born between the late 1990s and early 2000s, is leading the rebellion. As they spent formative years immersed in infinite scrolls and algorithmic feeds of social media, they have redefined content consumption habits — and with it, the rules of advertising.

In their world, “googling” information or products isn’t cool, so they rarely encounter search ads or display banners. TV and the newspapers are irrelevant too — so are the traditional commercials that come with them.

Even Google admits that nearly 40% of young people now turn to social platforms for discovery purposes, instead of Search or Maps.

The shift toward social platforms is especially evident in emerging markets like Latin America, the Middle East, as well as Central and Southeast Asia, where the median age is under 30 and smartphone penetration far outpaces desktop usage.

What’s crazy is that even social media platforms are now splitting between the cool and the quote-unquote lame ones. While TikTok, YouTube, Twitch, belong to the first category, Facebook is already giving ancient vibes. Instagram, too, is risking joining the second group soon, as reports show the app is slowly losing its appeal among younger audiences.

Then there comes Gen Alpha — the first AI-native generation — of whom we have yet to know what will influence their purchasing decisions. What we do know thus far is how they access information: through AI-powered chatbots and voice assistants.

In the ecosystems that Gen Z and Gen Alpha inhabit, intent-based ads no longer work. They’re replaced by content that appears in a scroll, in a reply, or in a conversation with an AI assistant.

Fatigue with ads and polished promises

The erosion of advertising doesn’t just end up with techno shifts. For modern youth, the resentment toward ads runs deeper, down to their economic reality.

Like any generation prior, Gen Z-iers aspire to homeownership, stable careers, relationships, and, maybe, even kids. Yet many see those things as financially unreachable.

Harris Poll for Bloomberg found that 45% of Americans aged 18–29 lived at home with their families in 2023 because they couldn’t afford to buy or rent their own space. That’s an 80-year high. And that tendency is becoming more common across the globe too.

In Gen Z’s view, wages haven’t kept up with housing costs, job markets are unstable, and long-term security is borderline fantasy. When every dollar counts, traditional ads that push consumption come off as detached and even insulting.

In contrast, the youth gravitates toward authentic content. Research by Ernst & Young showed that 92% of Gen Z-iers value being “authentic and true to oneself” as important. This quality determines which brands they trust, which accounts they follow, and ultimately, what they buy.

Hence, traditional advertising, with its clear selling intent, transactional tone, and overly polished offers, doesn’t quite land with this generation.

What happens when ads are replaced by influencers? Labubu

When ads fail to convince or even reach the young tech-savvy audiences, influencers come to the rescue.

”Influencers are everything.[…] I don’t care if you’re a bank, health care, anything — everyone is working with influencers,” declared Craig Brommers, the CMO of American Eagle Outfitters, at this year’s Cannes Lions.

He’s not wrong. More than that, the entire festival revolved around creator marketing — the model where brands leverage the credibility and reach of social media personalities to sell, persuade, and influence brand perception. In the era of eroding television ratings, traditional search abandonment, and demand for authenticity, influencer marketing has become every brand’s lifeline.

A recent exemplar of influencers’ power is the Labubu phenomenon. These quirky, tooth-baring elves made Pop Mart CEO Wang Ning a dollar billionaire — not a single ad involved. The dolls went viral after subtle endorsements from global stars like Blackpink’s Lisa and Rihanna. TikTok and Instagram creators then picked up the rest with unboxing trends, trades, and fan content.

Of course, Labubu’s “blind box” selling feature did pour fuel on the social media fire. But at the end of the day, it’s the creators and everyday people who boosted the trend with their content. Ads did nothing.

That’s the reality every brand is now up against.

The new-age advertising

Throughout its history, advertising has undergone multiple revolutions in its medium. But this time, what’s evolving is the purpose.

Beyond just selling, ads now have to become that voluntarily consumed content. Sometimes that can be achieved with an influencer casually involving a product in an “accidentally” viral TikTok. Other times, it’s the result of high-precision targeting through solutions like Retail Media or ad formats built into tools we use, like digital maps or streaming apps.

Some brands are already exploring these more subtle expressions of influence, like Toyota UAE. In their campaign Journey into Her Story, they used Yango Maps to honor the women behind the country’s street names through geo-triggered audio stories voiced by Emirati artists. Instead of pushing a message about cars, they contributed something lasting to the society they serve.

While that might be the death of advertising as we know it, it’s the birth of more enduring communication.

Advertising is dead. At least the way we traditionally think of it was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

Contextual targeting has had a wild ride since its lowly beginnings as a print-based advertising strategy. It went digital, gained popularity, and then sank to second place behind cookie-based targeting.

Now, it’s back in the limelight as third-party cookies grow scarcer. Privacy concerns are peaking, and users and advertisers alike demand more privacy-friendly ads.

Contextual targeting meets this demand by tracking webpage content instead of tracking individual users across the web. Advances in machine learning (ML) are making this approach increasingly effective. Context is back, baby — and it’s an exciting alternative to cookie-based advertising.

What is contextual targeting?

Contextual targeting is an ad targeting strategy that chooses which ads to place based on the content of a webpage. The goal is to show the user ads that are relevant to what they’re reading or viewing. This means if you’re reading an article about hiking, you might see ads for outdoor gear.

Instead of relying on third-party data, contextual targeting tools analyze keywords, topics, and even meaning to match ads with the right audience. It’s a smart way to reach interested users without intruding on their privacy.

Behavioral vs. contextual targeting

Behavioral targeting relies on users’ browsing history and personal data to show ads tailored to their past behavior. This includes retargeting, a strategy that shows ads to users who have visited a site or shown interest in a product before.

In contrast, contextual targeting focuses on a webpage’s current content. Ads shown are relevant to whatever a user is viewing, instead of a user’s behavior.

Not too long ago, the ad industry was shaken by Google’s announcement to phase out third-party cookies in Chrome, a move that sparked a race for alternative targeting solutions. While Google has since adjusted its plans and decided to keep third-party cookies in place, the conversation around privacy-first targeting hasn’t disappeared. Many advertisers and publishers still see the need to diversify beyond cookies, driven by growing user expectations, regulatory pressure, and the desire to future-proof their strategies.

In this context, contextual targeting retains a unique advantage — it was born cookieless.

How does contextual targeting work?

Contextual targeting works in several key steps:

1. Content analysis

To start, the ad platform scans the content of a webpage to understand its context. It analyzes text and images to identify keywords, topics, and themes.

2. Enrichment and ad matching

Once the content is analyzed, one of two things happens.

The old-fashioned approach, which is still used by some, is to categorize content into predefined topics and subtopics. For example, an article about shoes for summer might be categorized under “fashion” and “footwear.”

The innovative approach is to gather the content keywords, send them directly to an ML model without aggregation, and then have the ML model enrich them with synonyms and additional context features.

This method allows for a more holistic understanding of a page’s context, since it’s not limited to predetermined categories.

Once the keywords are enriched, the ML model selects the ads that best match the context and gets to work on placing those ads.

3. Real-time bidding (RTB)

Contextual targeting is often integrated into RTB platforms. This means that when a user loads a webpage, a milliseconds-long auction takes place to determine which ad to display based on the page’s content.

4. Monitoring and optimization

The effectiveness of the ads is continuously monitored. In AdTech platforms like Yango Ads Space, ML algorithms constantly analyze performance data to refine targeting parameters. This improves ad relevance and engagement over time.

When context and session data meet

Among the latest contextual innovations is the integration of anonymous session data. When you combine context signals with session data, magic (or close to it) happens.

I’ll use our contextual targeting algorithm at Yango Ads Space as an example. Suppose that, using anonymous session data from an advertiser’s site, the algorithm detects that a particular phone is often purchased during sessions that entered from an external phone review site. The algorithm then learns to place ads for that phone on the external review site in the future.

This enables ads to be contextually targeted and hyper-relevant to users, all while prioritizing privacy.

Diamonds are an advertiser’s best friend

From newspaper ads to ML-powered digital placements, it’s been a long road for contextual targeting. Now, third-party cookies are waning — whether or not the biggest tech names put the axe on them from the top down. The pressure’s on for cookieless targeting solutions.

Luckily, diamonds are made under pressure. Thanks to the renewed attention it’s receiving, contextual targeting has made light-year leaps in the past year alone. So stay tuned — we predict that the context renaissance has only just begun.

The context renaissance: How machine learning is making ads smarter without cookies was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

Users of various apps often take screenshots to capture content that is somehow important to them: they may want to share it (most likely via messenger or on social media) or save it to review later.

This turns a screenshot into a valuable signal of a high-intent interaction point, which can inform product logic or personalization strategies, be leveraged to improve user experience, and key in-app metrics such as retention, conversion, or content sharing.

Yet, many applications don’t take advantage of this signal to provide additional value to their users. In fact, most apps simply allow you to take screenshots without either notifications or restrictions.

When do users capture screens

To understand real-life motivation, I asked some friends and colleagues to share their recent examples of making screenshots, and grouped them into four patterns.

• To save important info they might need later, like promo codes, booking details, or event tickets.
• To share something quickly: from products and places to memes and posts.
• To remember things, price comparisons, quotes, design inspiration, or personal milestones.
• To compare options across apps or platforms, whether it’s hotels, real estate, or flight deals.

Each of these moments reflects high user intent. So, what if, with minimal behavior analysis or testing, apps could reliably anticipate what users want to do next and build flows to support it? A passive moment can turn into real engagement — a screenshot-driven interaction.

Screenshot interactions: a hidden opportunity for user engagement and app growth

The current mobile landscape still offers few strong implementations of screenshot-driven interactions, but the shift has already started. Leading platforms like Amazon, Farfetch, LinkedIn, and Google Maps are pioneering this behavior by recognizing when a screenshot is taken and immediately suggesting useful actions, typically related to sharing or saving.

They seem to be the first to recognize a critical insight: social behavior isn’t limited to social networks.

Well, indeed, mobile commerce is becoming inherently social: users browse, decide, and consult with others — all within the same device. On mobile, the gap between product discovery and sharing is minimal, while a screenshot often replaces a “save for later” or “ask for advice” button.

This creates an opportunity for e-commerce apps to step in. If the app detects a screenshot and surfaces a context-aware menu, such as options to share, save an item, or share a referral link, it meets users exactly at the point of highest intent.

This mechanic can serve loyalty goals, too. Imagine a user taking a screenshot of a product they’re considering. The app could respond with a push or in-app card: “Want to come back to this later? Add to your wishlist” or “Invite a friend to view and get 10% off”. These small nudges respect user intent and transform a passive act into an active conversion opportunity.

Let’s talk metrics: screenshots as a growth lever

At its core, screenshot recognition streamlines the user experience. By surfacing context-aware actions right when a screenshot is taken, apps reduce friction and help users stay focused on their intent. But app developers should also consider metrics far beyond the UX.

User attraction. An option to share the link directly turns a screenshot action into a lightweight referral mechanism, increasing the chances of attracting new users or re-engaging existing ones. Moreover, empowered with a built-in referral bonus, apps can encourage word-of-mouth in a way that feels natural and user-driven.

Conversions. When, after a screenshot, users are provided with shortcuts like “Add to favorites” or “Save for later”, it increases the chance that they will return to complete the purchase. This means less drop-off and higher average check size, as users are more likely to return to what they saved.

The “Share” option also creates extra conversion opportunities, as people they share with might remind them to buy more, or even visit the app and make purchases themselves.

Retention. Instead of disappearing into the abyss of their photo gallery, actionable screenshots help keep important in-app content within the app’s ecosystem — bookmarked, categorized, or nudging users to return. This reduces content loss and increases session return rates.

It can also increase session depth and time spent by helping users stay in context and reducing cognitive load, resulting in longer, more meaningful sessions. Screenshot-driven flows can guide users deeper into the app, jumping from discovery to favorites to checkout, all in one continuous experience.

Re-engagement. Screenshots can be used to trigger personalized incentives, such as referral codes, limited-time discounts, or follow-up nudges based on what the user captures. These contextual nudges help bring users back into the app with high relevance.

Feature adoption. Recognizing screenshots can also be a way to spotlight lesser-known features. For example, if a user captures a payment confirmation, the app could offer an option to save the invoice or set up recurring payments. This proactive behavior nudges users to explore functionality they may have otherwise missed.

Behavioral insights and data quality. As said before, screenshots are high-intent signals that reflect what users care about. Aggregated, anonymized data on what people screenshot most often can offer valuable UX and product insights, helping teams prioritize improvements, identify friction points, or discover what truly drives engagement.

Measuring the impact

All the metrics outlined above can be tracked easily with mobile analytics platforms like Google Analytics for Firebase and AppAnalytics by Yango. Both are free to use, support custom events, and offer developers full flexibility when it comes to integrating event-based tracking into their apps.

Firebase is used by many teams as their primary tool. However, complementing it with AppAnalytics can add an extra layer of depth: it is known for its near real-time data refresh, raw data export, and flexibility in creating custom reports without vendor lock-in. This makes it particularly useful for fast iteration cycles or when more granular, exploratory data cuts are needed.

Using both platforms can give a richer, more nuanced view of how screenshot-driven actions influence the user journey, helping validate hypotheses faster and optimize flows based on live behavior.

Screenshots and privacy: where is the line?

Despite the obvious benefits of screenshot-based mechanics across industries, some applications, mostly banking and financial, still block screenshots entirely, citing privacy and security concerns. Screenshots may contain personal data such as names, email addresses, photographs, or other identifiable information.

However, within the mechanics described, the screenshot file remains entirely on the user’s device and is not transmitted to the app’s servers or any third party. There is no background recognition of the image or analysis of the data contained within the screenshot. The app only reacts to the fact that a screenshot event occurred — a system-level signal that contains no personal data.

Even in the foresaid banking and finance practices vary: some apps block screenshots, others don’t, reflecting design choices rather than strict legal requirements. In retail and e-commerce, such restrictions often feel disproportionate. Product listings rarely contain sensitive data, yet disabling screenshots disrupts natural shopping habits like saving or sharing options, creating unnecessary friction and lost conversion moments.

Final thoughts

In a world where retailers compete for fractions of margin and fleeting user attention, even small behavior patterns can become decisive growth levers. While UI / UX teams are squeezing value from every pixel and every user gesture, screenshots are one of those underutilized signals.

Actionable screenshots are an overlooked surface for conversion, retention, and viral sharing. And when multiplied by millions of users, it can shape real business outcomes: more returning customers, more purchases completed, more reasons to stay inside an app ecosystem.

The all-or-nothing approach to screenshots is no longer the safest solution, let alone the most user-friendly one. Privacy and usability shouldn’t be at odds. With the right implementation, apps can protect sensitive data and respect users’ need to capture and share.

In this market, you don’t grow by guarding catalogs — you grow by meeting users where their intent is. And sometimes, that intent looks like a screenshot. In 2025, ignoring signals like this isn’t conservative. It’s leaving money on the table.

Screenshots are the new Clicks: apps can turn moments of interest into user actions was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

Retail Media is having a moment — and it’s only getting bigger. In 2025, it’s set to grow twice as fast as overall ad spend, per IAB’s digital advertising report. And by 2028, it will overtake Open Web advertising with $101 billion in revenue. In the Middle East, brands are catching on fast too; ad spending in the region is projected to rise by six percentage points this year.

But while everyone’s talking about it, not everyone’s sure what to make of it. Will it drive customers away? Slow down your site? Is it just another passing trend?

What is Retail Media?

Retail Media is advertising placed inside a retailer’s ecosystem, but not by the retailer itself. Instead, brands pay to show their ads on a retailer’s website, app, or even in-store screens. Think of it like the digital version of prime shelf space in a physical store, where brands pay you for visibility.

For instance, a sneaker brand can pay to appear at the top of a retailer’s search results or in a recommendation slot when a shopper views sports gear.

Retail Media isn’t new. What’s new is the scale, the tech, and the undeniable shift in how brands and retailers are approaching digital advertising. In effect,

• Retailers are no longer just selling products — they are turning their platforms into ad spaces that generate extra revenue.
• Brands aren’t chasing customers across different websites — they are reaching them exactly where purchases happen.
• What started as simple banner ads is now powered by smarter automation, first-party targeting, and data-driven recommendations.

Why Retail Media is on the rise

Traditional digital ads rely on third-party data from cookies, which is harder to track due to privacy rules and browser restrictions. As a result, targeting is weaker, ad costs are rising, and brands have less control over who actually sees their ads.

Retail Media does not have this problem.

It runs on first-party data, including hard IDs like emails, loyalty accounts, and purchase history. These identifiers are tied to real shoppers, not anonymous web activity.

Every time a customer searches for a product, adds something to their cart, or makes a purchase, those choices help shape a more relevant shopping experience. Advertisers can use these insights to show ads that match what people are already interested in, instead of randomly guessing and potentially pissing off their customers.

This closed-loop system benefits everyone. Advertisers run smarter campaigns, retailers open a new revenue stream, and customers see ads that actually align with their interests.

Who can use Retail Media?

Not every business can use Retail Media, but for those that can, the payoff is huge. Large ecommerce platforms, pharmacies, fintech apps, insurance providers, delivery services, and hospitality brands have the traffic and first-party data to make it work.

Some businesses, like many consumer packaged goods (CPG) brands, do not have direct-to-consumer ecommerce platforms. Instead of running their own ad networks, they can use Retail Media to place ads where it matters — on retailer sites that sell their products or on platforms where shoppers are browsing for related items.

Retail media works because it delivers value to everyone in the advertising ecosystem — brands, retailers, and even consumers.

For advertisers

• Better timing. Ads appear while people search for products, compare options, or add items to their cart.
• Better targeting. Retailer data allows brands to serve ads based on actual shopping behavior, instead of blind guesses.
• Clearer results. Every ad can be linked directly to clicks, purchases, and revenue, which makes ROI easier to measure.

For retailers

• New revenue stream. Retail media turns website traffic into a profitable ad business.
• Better customer experience. Ads are relevant to shoppers and do not disrupt their buying journey.
• Stronger partnerships with brands. Retailers attract more ad spend by offering high-quality placements with valuable shopper insights.

For consumers

• More relevant ads. Promotions match real interests and show products that fit their shopping habits.
• Smoother shopping experience. Sponsored listings blend naturally into product searches instead of interrupting them.
• Personalized deals. Retail media makes it easier for shoppers to find discounts and recommendations based on their preferences.

The biggest challenges in Retail Media today

Like any new opportunity, Retail Media has its hurdles. It is not all smooth sailing. Many retailers are still figuring out how to fit ads into their platforms without affecting the shopping experience. Advertisers, on the other hand, are cautious about moving budgets away from more familiar channels. The interest is there, but there are still a few key challenges. Chief among which are these three:

Retailers and advertisers are working in silos

Retailers and advertisers often use different systems, which makes it hard to run seamless campaigns across multiple sites and apps. Many companies are still figuring out how to integrate Retail Media into their existing tools. And unlike social media or search ads, where brands can run unified campaigns, Retail Media still lacks that level of seamless coordination. Organizations like IAB have established various committees and working groups and are regularly releasing guidelines and research to help the industry move forward.

No universal way to measure success

Unlike traditional digital ads, there is no single standard yet for tracking performance. What counts as a strong campaign on one retailer’s site might be measured completely differently on another. Without consistency, advertisers will struggle to compare results across different Retail Media networks. According to a 2024 survey conducted by BCG, 41% of brand and agency marketers ask for more benchmarks of their performance versus competitors. Experts from Nielsen also advocate for transparency and independent measurement of retail media attribution, warning against the so-called ‘walled gardens’.

Ad costs are climbing

As more brands invest in Retail Media, competition for ad space is driving up prices. Smaller businesses may find it harder to compete with companies that can afford higher bids for premium placements.

But this is not a dead end — it is an adjustment period

As Retail Media matures, more retailers are investing in better ad tech, AI-driven reporting, and cross-platform solutions.

Every new ad model takes time to settle, and Retail Media is no different. But solutions are already taking shape. More retailers are introducing unified dashboards, AI is making campaign reporting faster and more accurate, and competition is pushing platforms to become more transparent.

Brands don’t have to figure it out alone. As AdTech providers, we should make it easier for advertisers to seamlessly plan, launch, and measure campaigns across different retail platforms. Our job is to simplify campaign management, bring more transparency to performance tracking, and help businesses compete in this fast-growing space without breaking the bank. The trick is not to wait until everything is perfect, but to start now and build the advantage while the industry is still taking shape.

Retail Media: Deсoding the next big thing in digital advertising was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.

What’s next for the gaming industry? How do game developers keep up with rapidly changing technologies, shifting player behavior, and new monetization models? In this interview, we dive into these pressing questions with Gökhan Üzmez, a Game Tech professional and a top voice in the video game industry, who’s shaping the future of gaming.

Polina Invzhevatova, Business Development Manager at Yango Ads App Analytics, takes us through a deep conversation with Gökhan, where he shares his views on the latest trends, emerging technologies, and the evolving landscape of game monetization.

Whether you’re a game developer, an AdTech professional, or a monetization strategist, this interview offers a fascinating look at where the industry is headed.

Polina: The first question I’d like to ask is about how modern hybrid monetization models are balancing user experience with the need for revenue. How have these models evolved in recent years?

Gökhan: Back then, players used to watch interstitial ads while playing hyper-casual games; they rarely watched rewarded ads. But developers wanted to push them towards rewarded ads because that’s how games actually make money. Players don’t want to spend too much, but they also don’t want to watch ads; among other things, because an occasional interstitial ad pop-up kills the whole user experience.

If no one is making purchases or watching rewarded ads, developers earn nothing.

Nowadays, monetization looks a little more aggressive, especially with games that are scaling after a soft launch. But now, monetization is more segmented for each player, like:

Hey, do you want to spend as little as possible? We get that. But what do you want to spend it on? Is it content? Is it boosters? There are offers like ‘no ads for 3 days for just $1’. That’s a good deal because — let’s be honest — we know you’re not coming back on day 4.

Polina: With all these new regulations, how can developers still get the insights they need for monetization?

Gökhan: The weirdest thing is that despite strict regulations, the data is already out there anyway. Almost everyone has some type of social media account where they willingly share a lot — pictures, personal details, everything.

With games, it’s different. Let’s say you just logged in, and the game asks: Can we do this? Can we do that? Most players instinctively say: No, no, no! But if you say Yes, you’ll see ads with customized offers. If you help developers understand you, they can improve your experience.

In my opinion, this should be a win-win. The only way to adapt is to keep reminding players that developers are the good guys and that their data is in safe hands.

Polina: Have you seen any ad formats that feel almost too creative or deceptive?

Gökhan: First, there’s an interesting thing I’ve noticed, particularly in survival games. People can be lured to the game with, so to speak, “fake” gameplay ads.

What happens is people see video ads of the game with hyper-casual or fast gameplay, they like it, and download the game. But in reality, the game doesn’t offer that same gameplay as in the ads. However, players might still stick to it. That’s encouraging more video game companies to make more ads like that.

Second, there are those playable ads. Developers are looking at the most played games and incorporating their gameplay into playable ads. You know, that ‘click here to start moving this truck’ type.

And people fall for it — myself included. Just last week, I spent 20 minutes playing a playable ad just to see how it would end.

However, with playable ads, you don’t need to download the app, and there’s no call to action. So I can’t tell if it’s brilliant or absolutely crazy.

Developers are already making money by selling those ad spaces. They’re thinking: Let’s show an ad in an ad, in another ad. Kind of like that ‘Inception’ movie. And at some point, the users are probably going to forget where they are — in a gameplay ad or a real game. But it won’t matter as long as money is being made.

Polina: What’s the biggest trend you’re seeing in player habits right now?

Gökhan: The big deal is that people now have a shorter attention span, so it’s harder to keep them engaged for too long. Playtime per session is at an all-time low. What used to be 30–35 minutes per playtime session is not a thing anymore, and users can hardly commit to anything longer than 20 minutes.

In the era of shorts and tiktoks, players are looking for a quick adrenaline rush.

However, this also means that players are getting the fun out of the game quicker. And game developers can monetize more efficiently at a faster rate.

This creates more opportunities for new games and deeper content, as retention rates remain stable in many successful titles.

Polina: In what ways are generational changes influencing game development trends?

Gökhan: Developers are now targeting Gen Alpha (those born between 2010 and 2024), especially with games like Roblox or Fortnite.

Alphas are all about quick engagement — they want to jump in, have fun, and move on. They play games across multiple devices: phones, iPads, consoles, and computers. The demand for access anytime, anywhere is changing the way mobile games are developed.

So that’s changing the development of mobile games because they need to be accessible from different devices.

Also, this generation, born with an iPad in hand, sometimes shows less interest in face-to-face social interaction compared to older generations. Instead, they place a premium on socializing digitally.

To scale successfully with Gen Alpha, developers should focus on social features as well as cross-platform accessibility.

Polina: What emerging technologies do you think will make it possible for games to adjust dynamically to user behavior and improve their monetization?

Gökhan: I’m thinking of user-adapted content, but it’s a hard topic, technology-wise. Right now, developers make changes iteratively: they look at the data, adjust the gameplay, release the feature, and look at the data again. This way it is impossible to generate more content or develop a new feature immediately, not even mention personalize it.

However, in the future, I think games will be adjusting to how you play them. Let’s say you’re quick in an action or a puzzle game; it’s going to get harder as you progress, but if you’re not, it’s going to adjust so that it becomes easier. You won’t even notice it. Developers will aim to maintain a consistent level of fun to keep players engaged. Call it data science or an AI, or any other buzzword.

Another significant shift could come with the replacement of smartphones by VR and AR. Right now, we’re all depending on our smartphones. You can have an extra to it, like a smartwatch, but you can’t have just the smartwatch. Likewise, it is currently impossible to use solely an AR or VR tool.

I was the very first person to bring a VR Oculus to my country from the US in 2014 or 2015. It arrived as a prototype in a huge box, and I thought: “This is never gonna be a thing. It’s just too big.”

Remember, several years ago, Pokémon Go came out, where you needed to use your camera to find virtual creatures? This is an AR, but it is still limited to the phone. So I think, although all AR and VR tools are getting smaller and smaller, they won’t be daily used things unless we completely turn down smartphones to VR.

As the gaming industry continues to evolve, developers are adapting to new challenges and opportunities, from innovative monetization strategies to emerging technologies like AI and AR. The future of mobile games promises to be dynamic, with personalized experiences and multi-platform interactions becoming the norm.

Stay tuned as these trends reshape the gaming landscape — if you want to stay ahead of the curve, make sure to follow us for more insights from the industry.

“Let’s show them an ad in an ad, in another ad” — Gökhan Üzmez on the future of game monetization was originally published in Business Beat on Medium, where people are continuing the conversation by highlighting and responding to this story.