In an iGaming Real Talk interview in Dubai, the firm says the tool reads player emotion and context to guide operator actions.

SoftConstruct AI has unveiled RecSys, an AI Game Recommendation System, during an exclusive iGaming Real Talk interview recorded at AIBC Eurasia in Dubai on Thursday 30th April.

Mushegh Khachatryan, Chief AI Officer at SoftConstruct AI, said RecSys is designed to move beyond traditional recommendation models by interpreting player emotions in real time and accounting for context, with the goal of suggesting “the next best action” for operators. “You can understand your customer’s emotions in real time and suggest the next best action. We are building intelligent systems which can reason and act.”

Khachatryan said SoftConstruct built an AI Center of Excellence by hiring talent from outside the iGaming sector, and described RecSys as part of “production-ready agentic AI” intended to support personalised campaigns and decision automation. Surya Palli, host of iGaming Real Talk, said: “SoftConstruct is essentially building a Netflix-style personalised experience for the iGaming industry, where every player gets a lobby made just for them.”

Responsible gaming was also discussed, with the company claiming AI can detect risky behaviour faster and more consistently than human teams, and recommend timely breaks while balancing player protection with sustainable growth.

Khachatryan also stressed the need for explainable, controlled deployment. “AI should help teams perform five times better rather than replace them… Without proper boundaries, short-term momentum boosts with AI can actually hurt your company in the long term.” The company said operators can manage campaigns, personalisation and risk through chat interfaces, with at least 85% accuracy “from the first interactions,” and directed viewers to the full interview on the iGaming Real Talk YouTube channel.

The post SoftConstruct unveils RecSys AI game recommendation system at AIBC Eurasia appeared first on Eastern European Gaming | Global iGaming & Tech Intelligence Hub.

Traffy, a performance marketing agency specializing in in-app traffic, together with Traffy media buyers, prepared this material to share practical insights, testing experience, and scaling strategies in the iGaming vertical.

iGaming Campaign Overview: Scaling Strategy and CPI Optimization

In the gambling vertical, the Traffy team faced a typical growth challenge: how to scale campaigns without losing control over CPI and without sacrificing traffic quality.
The main goal was to reduce CPI while maintaining conversion rates to registration and deposit. Special attention was given to the scaling phase and subsequent campaign relaunches, where performance most often begins to degrade.

Before implementing the new strategy, single-format, unique gameplay creatives were used. While effective initially, this approach started to limit scaling potential as volumes increased.

How to Scale iGaming Campaigns Without Increasing CPI

As advertiser conditions allowed for active scaling, the task was to reduce CPI in order to improve conversions and lower deposit costs on the offer. CPI is unstable when using single-format creatives — it increases when scaling active ad accounts. If you rely on only one format, there is no real optimization, which leads to faster creative burnout and makes it harder to find new references that perform as well as previous ones.

Intraday dynamics were quite illustrative. In the morning, CPI could increase by up to 100%, then partially stabilize by midday, and rise again in the evening, adding around 20–25% to daily values. At the same time, campaigns continued delivering — they didn’t break, but efficiency gradually declined.

The main reason was not classic audience saturation. This was not a typical scale → saturation → CPI growth scenario. The problem was deeper: the single-format approach itself was limiting the algorithm. Creatives burned out faster, and the optimization system became less effective due to the lack of alternative signals.

Format burnout also occurred because there are limited in-app ad intelligence tools, and only a small number of teams have access to them due to their high cost. As a result, the overall pool of proven working creatives is relatively small. When only one creative format is used, it quickly burns out, forcing the search for new references — but often there is nothing left to scale further.

Why Single Creatives Formats Lead to Higher CPI and Instability

During testing, the team reached an important insight: CPI growth and instability were directly linked to using only one type of creative.

In practice, the algorithm was constrained by a single format — it had no ability to redistribute budget or optimization signals. This increased CPI volatility and accelerated burnout.

This understanding came empirically through a series of tests. It also became clear that some formats, such as news-style creatives and playable formats, were underutilized, which limited scaling potential.

Creatives Testing Strategy to Reduce CPI in iGaming Campaigns

The team shifted focus toward creatives diversification.
They decided to run multiple creatives types simultaneously, giving the algorithm more flexibility. Three key formats were introduced:

• interview
• gameplay
• social gameplay

The main share of traffic was allocated to interview creatives (50%) with a CPI of about $8.58, while gameplay accounted for 25% with a CPI of ~ $6.5, and social gameplay — the remaining 25% with a CPI of ~ $10.23 (Moloco Ads).

Scaling was primarily achieved through relaunching campaigns with increased budgets.
The team tested 4 to 6 creatives per campaign with a $250 budget. If performance was strong, the campaign was relaunched with a higher budget, and its behavior was closely monitored.

How campaigns were launched:
If new campaigns included only new creatives, they were kept and scaled alongside existing active campaigns, or relaunched when needed.

Testing was conducted in parallel: different formats were launched simultaneously, and the decision-making cycle took 2–3 days. The primary evaluation metrics were CPI and cost per registration, followed by CPA.

Results: Lower CPI and More Stable Campaign Scaling

Switching to a multi-format strategy delivered a noticeable impact. Previously, overall CPI ranged between $13–18, but after implementing the new approach, it decreased to approximately $8.4.

The key improvement was scalability: campaigns could be scaled without sharp CPI spikes. Although CPI growth with increased volume remains inevitable, it became much more controlled and predictable.

From a performance perspective, different formats showed varying CPI levels. The best result was achieved by social gameplay, while other formats demonstrated higher CPI. However, this did not reduce their importance in scaling.

For scaling, social-style formats and interview creatives performed best. Interview creatives tend to burn out faster, while social gameplay creatives — being bright, visually engaging, and highly informative — scale more effectively.

No format proved to be universally best for growth. All three formats scaled effectively, and their combination gave the algorithm more opportunities for optimization.

Best Practices for Launching and Scaling iGaming Creatives

In practice, the most effective approach is straightforward: launch 4–6 creatives across 2–3 formats (interviews, social approaches) and run them either sequentially or simultaneously, as done at Traffy.

A specific recommendation from a Traffy buyer: use playable elements in the end card — this can significantly reduce CPI. Creating a unique end card for each creative in the same style also lowers CPI and increases engagement with the creative itself.

At the same time, it is important to understand that there are no “dead” formats: over time, any creatives will burn out regardless of their type.

Conclusion: Multi-Format Creatives Strategy for Predictable iGaming Scaling

Creatives format diversification proved to be the key factor for stable scaling. Instead of trying to control CPI within a single format, the team expanded the optimization space. This allowed for better traffic redistribution, slower creative burnout, and reduced performance volatility.

As a result, scaling became not only more efficient but also more predictable — without the sharp CPI spikes typical of single-format strategies.

If you have any questions or would like to apply these strategies to your campaigns, feel free to submit a request via our website or reach out directly at [email protected].

The post In-App Creatives in iGaming: How to Lower CPI and Scale Campaigns with Moloco Ads appeared first on Eastern European Gaming | Global iGaming & Tech Intelligence Hub.

After spending 10+ years running your own casino brands, you develop a very specific frustration with the tools available on the market. You know exactly where the funnel breaks because you’ve watched it break on your own traffic. With your own money at risk.

That point is where the Kanggiten iGaming platform started. What would a platform look like if the people building it had already operated 50+ brands and managed over 3 million players?

It would look different. And it does. Here are five features that reflect that origin.

Modular Architecture with Conversion-Tested Components

Kanggiten is modular by design. Operators choose the components they need, whether that’s a casino engine, sportsbook, CRM, affiliate management, analytics, or game aggregation, and add more as their business grows. No platform migration, no rebuilding infrastructure from scratch. Each module is standalone, yet fully integrated when combined.

First, this removes platform lock-in. An operator can start with a Kanggiten white label iGaming platform setup and expand into a full turnkey model over time without switching providers. Second, every module ships with conversion logic already built in. Registration forms, deposit flows, lobby layouts, bonus mechanics. All of it has been A/B tested on live traffic across real projects.

Operators on the platform have reached registration-to-deposit conversion rates of up to 70% (results vary depending on geo and traffic source). That number comes from systematic UX research, continuous hypothesis testing, and a product team that treats every redundant click as lost revenue.

Retention as a Full Operating System

Nowadays, retention costs less than acquisition and scales more predictably. The harder question is how a platform supports it in practice. Most setups offer a bonus toggle here, a CRM email there. Kanggiten treats retention as a connected system that spans the entire player lifecycle.

That means onboarding sequences, CRM-driven reactivation flows, bonus economics with built-in anti-abuse logic, and weekly engagement mechanics like tournaments, prize wheels, and achievement systems. Layered, tested, and iterated based on live performance data.

Operators on the platform consistently reach retention rates of up to 39%, against an industry benchmark that typically sits between 30% and 35%.

“Retention is a system, not a single feature. It’s built on analytics, testing, and continuous optimization,” – says Ivan Korkin, Head of Account Management at Kanggiten. – “Our focus is on onboarding, CRM flows, bonus economics, and reactivation, because that’s where sustainable growth actually lives.”

Segmentation That Accounts for Geo and Behavior

Grouping players by deposit size or registration date only gets you so far. Kanggiten runs split tests across different traffic types and builds tailored monetization scenarios for each identified segment. The deeper the behavioral model, the more segments emerge. And this directly translates into the more precise revenue impact.

The geo-specific layer adds another dimension. The platform reflects local user behavior from the first interaction, including UI/UX patterns, number formatting conventions, payment preferences, and content expectations. Small details, but they compound.

“For example, in the Turkish market, users expect percentage values displayed before the number. Even that inconsistency can hurt conversion,” – Ivan explains. – “In LATAM, fraudulent registrations are a known issue, so we analyze behavioral patterns in those regions and apply AI-based verification. A localized product and a translated product are two very different things.”

AI Applied Across the Operations Stack

Kanggiten deploys AI only where it produces measurable operational gains: 

• fraud detection, 
• KYC and identity verification, 
• content generation, 
• customer support automation, 
• cost optimization. 

It works across the full stack, reinforcing every other module in the process.

In high-risk geos, AI-based smart verification reduces the burden on support teams and cuts fraud activity without adding friction for legitimate players. In content and support workflows, it accelerates turnaround and lowers cost per interaction.

Over 20 Communication Channels with Built-In Flexibility

It’s not a big secret that your channel reach has a direct impact on revenue. Kanggiten supports more than 20 active touchpoints, including Telegram, social platforms, push notifications, SMS, email, and webhooks. Operators get multiple paths to reach players where they already spend time.

This matters especially in markets where communication methods shift or get restricted without warning. Channel diversity at Kanggiten is part of the core infrastructure. When one path closes, others are already running.

Why It Adds Up

Together, these features make up a platform built around how operators actually work. The same thinking extends across the full ecosystem, including the Kanggiten Affiliate Platform (powered by FireAff) and the Kanggiten CRM & Marketing System (powered by InTarget).

And the industry has started to take notice. In March 2026, Kanggiten was named Best Live Casino Provider at the GamingTECH CEE Awards during the Hipther Prague Summit, an event that brought together 400+ attendees from 35+ countries.

Why? Because every feature was shaped by a team that has run brands, managed players, launched campaigns, and been held to the same KPIs their clients track every morning.

The post 5 Platform Features That Only an Operator Would Have Built: Kanggiten’s Perspective appeared first on Eastern European Gaming | Global iGaming & Tech Intelligence Hub.