Wishroll / Status: Cutting AI costs by >95%, scaling to 500K+ DAUs, and driving time spent per user to over 1.5 hours per day

Quick facts: Status by the numbers

Game concept: AI-powered social simulation game where players live as anyone they can imagine on a fully simulated social network platform.
Next-level player engagement: Average daily playtime of 1 hour and 36 minutes per user
500,000+ daily active users: Successfully scaled from limited beta to mass audience following February 2025 full release
Developer: WishRoll, a consumer mobile app company, previously built Kiwi, a music sharing app that had over 2 million downloads across the world. WishRoll's family of apps have been downloaded over 5 million times worldwide since 2021.
Explosive growth: From 100 to 190,000+ Discord members within days of limited beta
Top 10 lifestyle app: Maintained premium ranking in App Store
Top-grossing AI app: User-aligned monetization allowing continued growth

Game-changing results at a glance

95% reduction in AI costs: Cost reduction for a clear path to scale and profitability
Auto-scaling infrastructure to support millions of DAUs: Successfully scaled from limited beta to mass audience
Zero downtime: Consistent sub-second response times even at peak load
Enhanced user engagement: Continued growth and retention driving premium ranking in App Store
Stronger monetization and business model viability: Advanced user engagement and new feature investment driving remarkable monetization

Top engagement & retention: Ran tests on models, prompts, and architecture to increase engagement and retention to levels well beyond industry norm

Founded in 2020, Wishroll is a consumer app development company. When deciding on their next project, CEO and co-founder Fai Nur and her team were driven by a simple question: "What would have excited us as teenagers?"

That question led Wishroll to the idea for Status, a groundbreaking AI-powered game best described as The Sims for social media. In Status, players roleplay as anyone they imagine—whether it's a world-famous pop star, a fictional character, or contemporary Silicon Valley tech leader. The goal is to amass followers, develop relationships with other characters, and complete unique milestones.

When a player chooses Snow White, for example, Status dynamically generates social media profiles for the seven dwarves, the evil stepmother, Prince Charming, and even random villagers—all posting and interacting in character. Players then engage through in-character posts in a fully realized social media universe.

This vision struck an immediate chord with players, with TikTok buzz driving over 100,000 downloads during the limited beta, while the game's Discord community ballooned from a modest 100 users to 180,000 within a few days. But this remarkable success introduced critical technical and business challenges that threatened the game's viability at scale.

Technical Problems & Business Impact: The AI Scale Crisis

Wishroll's small five-person team faced an impossible dilemma: maintain quality and go bankrupt, or sacrifice quality and lose players. As CEO Fai Nur put it: "We're not ML Ops people or AI researchers—we're focused on the consumer experience. But scaling this kind of product was prohibitively expensive. We needed to dramatically lower costs while maintaining the quality expected by our users."

The holistic AI solution enabling massive-scale success

Rather than approaching each challenge in isolation, Inworld implemented a comprehensive strategy that addressed the interconnected nature of Wishroll's AI scaling challenges:

1. 95% cost reduction: Strategic AI architecture that scaled profitably

Inworld's comprehensive cost optimization didn't just switch to cheaper models—it reimagined the entire AI architecture:

Specialized prompt orchestration: Replaced single all-purpose prompts with three targeted prompt categories (character tweets, random tweets, news updates), dramatically improving both quality and efficiency
Custom evaluation metrics: Developed automated evaluators specifically measuring factors critical to Status (content diversity, emoji usage appropriateness, humor)
Load testing at production scale: Verified performance at over 500M tokens per minute and 2,500+ queries per second

RESULT: AI costs plummeted by ~95% while maintaining or improving content quality—transforming an existential financial threat into a sustainable business model.

"This collaboration has allowed us to take Status from a limited beta to a game that can reach millions of people," said Fai. "And it's not just about the present—Inworld is helping us stay ahead of AI advancements so we can keep optimizing and delivering the best experience for our users."

2. Real-time quality monitoring: From "Facebook Mom" posts to authentic content

Initially, using the models that meant the cost bar meant a 78% drop in engagement. Despite cost being a factor, maintaining the game's authentic voice and upholding user engagement was non-negotiable. Inworld’s products allowed comprehensive systems to ensure efficient content quality at scale:

Player-driven feedback loops: Leveraged in-game rating systems to improve AI-generated posts on accuracy, humor, and engagement
Comprehensive telemetry: Inworld enabled robust analytics to track quality metrics across different interactions
A/B testing framework: Systematically tested content variations against real player reactions to continuously refine the experience

"Our users don't want posts that sound like a 'Facebook mom,'" Fai explained. "They want cool Gen Z slang, not something cringey. Inworld's team iterates quickly based on that feedback, ensuring our users stay happy."

RESULT: Players experienced consistently high-quality, authentic interactions rather than the generic content that had previously driven away users in earlier tests.

3. Multi-provider freedom: Breaking free from AI vendor lock-in

Rather than tying Wishroll to a single provider's ecosystem, Inworld built a flexible infrastructure to leverage the best technologies:

Provider-agnostic architecture: Inworld products included abstraction layers allowing seamless switching between models from different providers
Rapid model evaluation: Enabled side-by-side comparison of new models as they were released, with automated quality assessment
Unified analytics dashboard: Created single source of truth for performance metrics across all providers

"With other providers, you're locked into their ecosystem," Fai said. "Inworld gives us flexibility. If a model doesn't work for our use case, we can test another one quickly without losing momentum."

RESULT: While competitors remained locked into increasingly outdated AI models, Status continuously incorporated cutting-edge advancements, maintaining its technological lead without disruptive migrations.

4. Production-grade infrastructure: Zero downtime at 500K+ users

Scaling reliably required sophisticated infrastructure designed specifically for Status's demanding requirements:

Comprehensive telemetry: Inworld suite supported production-grade observability with detailed metrics, traces, and logs for every AI interaction
Anomaly detection: Automated identification of quality or performance regressions
Custom dashboards: Real-time visibility into key performance and quality metrics

"The fact that Inworld has been so responsive and helpful helps us move faster — and the fact that they have such a deep understanding of so many different models that would work for our use case has been key, too," Fai shared.

RESULT: Status scaled seamlessly to over 500,000 daily active users with consistent sub-second performance—even during viral growth spikes that would have crashed standard implementations.

5. Data-driven decision making: Complete visibility into AI performance

Before Inworld, Wishroll had limited visibility into their AI systems' performance. Inworld deployed:

Automated AB testing and experimentation: Using performance, usage and cost thresholds to automatically switch logic and models
Predictive load management: Built systems to anticipate and handle traffic spikes without degradation
Multi-layered fallback logic: Designed architecture that automatically routed around provider limitations

RESULT: Wishroll gained complete visibility into their AI systems, allowing data-driven optimization and rapid problem resolution.

6. Controlled evolution: AI that gets better every day – On cost, quality and user-personalized engagement

Most critically, Inworld’s product suite enabled a durable framework to ensure Status's AI continuously improved rather than stagnating:

Learning service: Deployed systems to automatically improve based on player interactions and feedback
Continuous model evaluation: Established ongoing assessment of newly released models against Status's specific requirements
A/B testing at scale: Integrated Inworld to test improvements with real users before full deployment

"The real upside with Inworld is that it's a continued relationship that allows us to keep optimizing performance and also keep lowering the costs because now we have this team of ML experts working with us," explained Fai. "Our goal is to get this down to a few cents per user per day, and that's something that you can only do with true experts and an ML team that really understands the trade-offs and techniques for accomplishing that."

RESULT: While competing games quickly became repetitive and predictable, Status maintained fresh, engaging content that evolved with player preferences—driving exceptional retention and long-term engagement.

Transformed business outcomes: From existential threat to market leadership

Inworld's comprehensive solution transformed Status from a technically challenged concept into a thriving, sustainable game platform:

Financial transformation

Viable economics: Reduced per-user AI costs from bankrupting $12-15 to sustainable cents per day
Predictable scaling: Established cost structure that actually improves rather than deteriorates with scale
Engineering focus shift: Freed Wishroll's developers from constant MLOps troubleshooting to focus on new features

Player engagement metrics

Feature acceleration: Because they were able to lower costs and delegate ML issues, WishRoll was able to launch new features to keep the game fresh. These included a calendar system where users can roleplay live events like concerts or parties. This was critical to keep players engaged with Status and reduce churn but only possible because the AI cost savings allowed them to add more functionalities.
Consistent quality: Maintained authentic, engaging character interactions even at massive scale and with new universes.
Long-term retention: Prevented the "content fatigue" that typically plagues AI applications.

"Working with Inworld has been the most positive experience I've had with another company," said Fai. "They've been incredibly hands-on, responsive, and fast. It's like having an external ML team dedicated to helping us solve our problems."

Competitive advantage

Time-to-market acceleration: Launched full release months ahead of what would have been possible without Inworld
Operational excellence: Built infrastructure and systems that competitors simply cannot match
Future-proofed technology: Established framework for continuously incorporating the latest AI breakthroughs

Player impact: Real experiences

The Inworld difference: True partners, not just providers

Unlike Wishroll's previous experiences with major AI providers (where support requests took days or weeks), Inworld operated as a genuine extension of their team:

"When you work with external parties you never know what you're going to get, but Inworld has been really hands-on and fast with helping us," Fai explained. "It's honestly been the most positive and best experience I've ever had working with another company on a problem like this."

This hands-on approach was critical as Wishroll hustled to get Status out in a full release. The combination of deep technical expertise and responsive collaboration made all the difference in meeting their launch timeline.

Beyond Status: AI that scales without compromise

Wishroll's experience proves that AI-powered consumer applications can achieve massive scale without compromising quality or economics—when implemented with the right approach. Inworld now remains a partner for the future, continuing to improve all models and AI logic to ensure evolving quality, cost and user engagement.

For developers facing similar challenges, the lesson is clear: success requires more than just great models; it demands holistic systems thinking across costs, quality, flexibility, observability, and continuous evolution.

"This collaboration has allowed us to take Status from a beta with limited access to a game that can reach millions of people," Fai summarized. "And it's not just about the present—Inworld is helping us stay ahead of AI advancements so we can keep optimizing and delivering the best experience for our users."

Ready to scale your AI without breaking the bank?

AI doesn't have to be financially ruinous at scale. Inworld is solving the deep technical challenges that make production-grade AI viable for consumer applications, from cost optimization to continuous evolution.

Get started with InworldGet in touch to discuss signing up for the Inworld License.