Over the past year, AI-powered apps have entered a phase of rapid growth. According to data from AppGrowing, the number of active AI app campaigns in Q4 2024 rose by 20% compared to Q1.

While mature markets like North America and Europe remain key battlegrounds, emerging regions such as Southeast Asia are drawing increased attention for their vast untapped potential.

As competition intensifies, the question for developers is how to seize the opportunities these emerging markets present. One answer may lie in the strategies of local players. Vietnamese developer Xtech began exploring the AI space in late 2023 and has since made generative AI its strategic priority for 2024. The company recently launched PixArt: AI Photo Generator (hereafter referred to as PixArt), its flagship generative AI photo app.

In this exclusive interview, AppGrowing speaks with Minh Chau Pham, CMO, and The Anh Tran, CGO of Xtech, to explore the opportunities and challenges generative AI apps face in both the Vietnamese and global markets—from product design to market positioning and user acquisition strategy.

Part I: Company & Market Landscape

To start, could you briefly introduce Xtech to our readers?

CHAU: Established just last year, Xtech is a technology company dedicated to creating human-centric tech solutions. We've achieved early success with our health and utility apps.

This year, we're focusing on generative AI, launching products like an AI photo generator and an AI-powered home decoration app, expanding our offerings while maintaining our commitment to user-friendly innovation.

When did Xtech begin focusing on generative AI, and what opportunities did you identify in this space?

CHAU: We started exploring AI products at the end of last year, recognizing the explosive growth and demand in the generative AI category. What stood out to us was that while the trend was taking off, there were still relatively few strong players in AI products—especially in photo generators.

The creativity, personalization, and fun that AI photo tools offer made us optimistic about their viral and commercial potential. There’s a huge opportunity here—not just in art and entertainment, but in how people express identity and culture visually.

From your perspective, what stage is the generative AI market in Vietnam currently at?

THE ANH: I think Vietnam’s generative AI market is in early growth: competition is rising but not crowded, with local and international apps. The user base - mainly young, tech-savvy people - is expanding fast.

However, monetization is still developing. Most users are not fully used to paying for AI-generated content or subscriptions yet, so we’re experimenting freemium features and ads to understand what works best for local users. It’s promising, but the market is still maturing.

Beyond Vietnam, which regional markets are you prioritizing, and how do they differ from your home market?

THE ANH: Beside the local market, we’re focusing on Southeast Asia, North America and Europe. We’ve discussed mainland China but, due to strict regulations, fragmented app stores, and entrenched local giants like Tencent and ByteDance, we’ve decided to prioritize other regions for now.

Compared to Vietnam’s young, experimental audience and early-stage monetization, these markets feature mature payment habits, higher advertising costs, and more demanding localization requirements—so we’re tailoring our go-to-market strategy accordingly.

Part II: Product Strategy & User Acquisition

What inspired the development of PixArt? What sets it apart from other AI photo generator apps on the market?

THE ANH: We started with a simple insight: people crave fast, fun, personalized images without needing design skills. So we built a mobile-first app with a lightweight model that lets users generate high-quality photos in seconds.

But speed isn’t enough. We spent months fine-tuning our engine on regional aesthetics, so a filter in India feels distinctly different from one in the US. And instead of typing out detailed text prompts word by word, we built a one-tap preset gallery: you pick a style and watch your photo transform instantly.

Behind the scenes, our automated pipeline enabling new effects and filters roll out faster, while a smart on-cloud balance keeps performance smooth and server costs under control. We’ve also built in ethical safeguards—watermarks, content filters and clear policies—plus end-to-end localization of UI text and ad creatives. The result? A blend of speed, cultural resonance and trust that truly sets us apart.

In your view, what is the biggest challenge in launching a generative AI photo app today?

THE ANH: Honestly, the toughest part is cutting through the noise: with dozens of apps promising AI magic, acquiring users profitably and keeping them engaged is brutal. High marketing costs, rapid feature parity among competitors, and fickle user trends mean you have to constantly innovate—both in tech and UX—while controlling CPI.

On top of that, ethical and data-privacy concerns add layers of complexity. Balancing these demands—standing out, controlling costs, and moving faster than the competition—is really the toughest hill to climb.

When it comes to user acquisition, how do you maintain a competitive edge?

CHAU: With DeepArt, our AI photo generator, we took a data-driven and highly localized approach to user acquisition. We started by using AppGrowing’s insights to identify emerging opportunities in Tier 3 markets—specifically countries like India, Brazil, and parts of Southeast Asia. These regions showed strong demand for AI creative tools, yet relatively low saturation from global competitors. That was a key advantage for us.

From the beginning, we designed DeepArt to align with local user preferences. This meant more than just translating the app—we adapted the entire user experience. In-app content, onboarding flows, templates, and creative assets were all localized to match cultural nuances, popular styles, and user behavior patterns in each region. On the marketing side, we created region-specific ad creatives and tested them rigorously to understand what resonated best.

Once we gained traction and positive user signals in those markets—such as strong retention, organic growth, and positive feedback—we expanded into Tier 1 countries. But we didn’t just copy-paste our strategy. Instead, we repeated the localization process, tailoring both our app content and marketing strategy to fit the expectations of more mature markets in North America and Europe.

In short, localization has been the foundation of our competitive edge—from ASO to paid campaigns, and even product design. And tools like AppGrowing have helped us monitor the creative trends and campaign strategies of top-performing competitors in each region, so we’re able to adjust quickly and stay ahead. This approach allows us not only to scale efficiently, but to build long-term engagement and brand affinity across diverse markets.

How have third-party tools like AppGrowing helped you optimize your UA performance?

CHAU: We use AppGrowing daily as a core part of our user acquisition strategy. It helps us stay on top of market trends and closely monitor competitor activity, especially in terms of creative strategy. One of the key features we rely on is tracking ad creatives—what formats competitors are using, what messages are resonating, and how trends evolve over time.

In terms of metrics, we pay close attention to impressions, active days, and especially the Active Index, which gives a strong signal of how well a creative is performing over time—not just in reach but in sustained effectiveness. This helps us prioritize concepts with long-term potential.

We also closely monitor channel allocation. Understanding where our competitors are putting their budget—whether it’s Meta, TikTok, Google, or other networks—helps us benchmark our own strategy. For example, if we see a shift in channel focus for a certain geo or season, we can adjust quickly and test new placements before the market gets saturated.

Finally, we use AppGrowing’s Holiday calendar to plan timely, localized creatives, especially for markets with unique holidays or cultural events. It’s a key tool that supports both our marketing and product localization strategies.

Speaking of creative strategy, how do generative AI apps differ from other categories in terms of ad creatives? What are the key elements when producing these assets?

THE ANH: Generative AI really flips the script on traditional creatives because it isn’t just a static ad—it’s a live demo of what the tech can do. Instead of showing a single finished image, you’re selling the magic of instant, personalized transformations, so our visuals need to feel dynamic and playful.

When we craft creatives for this kind of product, we always spotlight that “wow” moment: a quick before-and-after, a tap yielding something totally unexpected. We keep copy short and snappy—users shouldn’t have to read paragraphs to get hooked.

Given our focus, we ensure our creatives resonate by showcasing diverse styles and faces that look familiar and appealing to our target regions. It’s not just about a cool image; it’s about a cool image that feels for them. We also lean into real-world use cases (such as travel selfies to professional portraits) so people can instantly imagine themselves using it.

What’s your take on the use of generative AI in performance marketing itself?

CHAU: I think generative AI is becoming a key tool in performance marketing—it’s no longer just a “nice-to-have,” it’s quickly becoming essential—especially for content creation and campaign testing. One of the biggest challenges in user acquisition is keeping creatives fresh and relevant across multiple regions and platforms. With generative AI, we can rapidly create variations of ad assets—different languages, formats, visuals—while still keeping the core message consistent. That kind of speed and flexibility used to take days or weeks and now happens almost instantly.

Part III: Looking Ahead

With AI technology evolving rapidly, which directions is Xtech paying close attention to?

THE ANH: The pace of change is just incredible.Definitely on-device AI and model optimization. Making these powerful models smaller, faster, and more efficient without sacrificing quality is a huge focus. The more we can do directly on the user's phone, the better the experience – faster results, better privacy.

Then there's everything happening with multimodal AI. The idea of models that can seamlessly understand and generate content across text, images, video, and even audio is mind-blowing. Imagine being able to describe a scene and not only get a picture but a short video clip with an accompanying soundtrack. That's the kind of stuff that gets us really excited about future possibilities.

And, of course, ethical AI and responsible development isn't just a buzzword for us; it's foundational. So, we're constantly tracking new techniques for bias detection and mitigation, content sources, and creating safer AI systems.

From an application-level perspective, which verticals within generative AI do you believe hold strong potential going forward?

THE ANH: Beyond photos, I'm particularly excited about AI video generation. We're seeing tools that can turn a few text prompts into short clips, but imagine when that becomes as fluid and high-quality as image generation is today.

Then there's the whole world of 3D asset generation. Think about game development, architectural visualization, or even creating assets for the metaverse. Being able to type "a futuristic armchair with glowing blue accents" and get a usable 3D model would be a game-changer for creators and designers.

Music and audio generation is fascinating too. We're seeing AI that can compose original scores, create custom sound effects, or even generate realistic voice-overs in multiple languages. This could revolutionize everything from indie game development to podcast production.

I'm also keeping a close eye on generative AI for personalized learning. Imagine educational content that adapts not just to your knowledge level but to your learning style, interests, and cultural context. A history lesson that's uniquely engaging to you because the AI knows how to present it in a way that clicks with your brain.

Conclusion

According to data from AppGrowing, global advertising for AI apps grew by 5% quarter-over-quarter in Q1 2025—signaling not only sustained market momentum but also intensifying competition.

Against this backdrop, Xtech has provided a compelling case study in how a data-driven and deeply localized strategy can be leveraged to build AI products that resonate across diverse markets.

At AppGrowing, we share this data-first philosophy. Our mission is to empower developers and marketers with in-depth creative analysis and comprehensive market intelligence, helping them localize more effectively and scale globally.

As the industry continues to evolve, we believe it’s the real-world experiences of developers that will shape its future. And only by deeply understanding their needs can we offer truly impactful support on the path to global expansion.

We warmly invite more developers to join the conversation—so together, we can drive the industry forward.