This blog was translated from Japanese to English, and was originally published here.
Asoview Co., Ltd., operator of "Asoview!", a leading leisure and experience booking platform in Japan, sought to build a fraud detection capability independent of its payment service providers (PSPs). To achieve this, the company implemented Sardine, a next-generation fraud detection solution.
DGBT spoke with Junya Ebe, Senior Executive Officer and CIO, about the path to that decision and the results that followed.
This article is based on a presentation delivered at "DGBT CommerceDay 2025," held in December 2025.
"A safe experience" as a core service principle
Can you start with an overview of Asoview and what guides how you run the service?
Asoview operates two interconnected products: "Asoview!," a consumer-facing platform for booking leisure and outdoor experiences, and "Urakata," a reservation and ticketing management SaaS for leisure facilities and activity operators. Together, they function as an integrated platform, currently serving approximately 6,000 partner businesses and 16 million registered members. Our core user base skews toward adults in their 20s to 40s, with strong representation among couples and families. The pandemic was an extremely difficult period, but we used it as a turning point by capturing growing demand from leisure facilities looking to digitize their operations.
The principle that guides everything we do is safety and security. When people go out for leisure, those moments need to be special. We cannot be the reason that experience is ruined.
In concrete terms, we treat the following three areas as the pillars of what "safety" means for our users:
- Active investment in information security and customer data management
- Prioritizing system availability and stability
- Ensuring seamless experiences for both users and facility operators
When high-demand tickets go on sale and traffic spikes sharply, transactions need to keep moving. That’s why we maintain backup systems for outage scenarios and treat availability as a constant priority.
Breaking away from PSP dependency
What led you to build your own fraud detection capability?
When you rely on a PSP's bundled fraud tools, every time you switch providers, your entire fraud prevention posture resets. The data you've accumulated, the patterns you've learned - none of it carries over. We concluded that having our security posture tied to the commercial decisions of a payment provider created a structural risk over the medium to long term.
Safety and security are foundational to what we offer. We decided to build an independent fraud detection capability, one that wouldn't be held hostage by whatever PSP we happened to be using.
Six evaluation criteria and why Sardine won
How did you approach the vendor evaluation?
We assessed candidates across six dimensions:
- Detection technology flexibility (AI vs. rules-based, methodology)
- Latency and throughput
- Ease of implementation and ongoing operation
- Flexibility in applying coverage to specific product categories
- Cost-effectiveness
- Long-term viability
Processing speed was the factor we weighted most heavily. When a popular ticket goes on sale, transaction volume can spike in an instant. Any latency in payment processing at that moment directly degrades the user experience. Sardine's average fraud response time of under 0.2 seconds, combined with their ability to handle hundreds of requests per second under peak load, was a decisive factor.
We also valued the ability to apply fraud detection selectively, targeting higher-risk product categories rather than applying blanket coverage across all transactions. That flexibility translated directly into a more rational cost structure.
How did you evaluate Sardine's technology specifically?
The combination of behavioral biometrics and device intelligence stood out as particularly effective. The system analyzes subtle behavioral signals such as mouse movement, typing cadence, and the tilt of a smartphone in the user's hand to distinguish bots from legitimate users. In an era where generative AI makes it increasingly easy to fabricate personal information, unconscious behavioral patterns are far harder to spoof. That's precisely where this detection approach has an edge.
Sardine's ability to identify the true origin of connections, even when IP addresses are manipulated to make overseas access appear domestic, also caught our attention. We see this type of spoofing frequently in Asoview's transaction data, and the impact of that capability has been tangible.
The hybrid architecture combining AI and rules-based logic was another point in Sardine's favor. Pure AI takes time to achieve high accuracy. Pure rules can't keep pace with evolving fraud tactics. The two-layer approach, where each compensates for the other's weaknesses, gives us confidence in production.
How did you assess Sardine's long-term trajectory?
Fraud tactics evolve constantly, so a solution's ability to keep pace matters enormously for a long-term commitment. Sardine ships new features on a monthly cadence, which signals sustained investment in the product. The recent addition of a scoring functionality to distinguish legitimate AI agents from malicious bots is a good example of that forward momentum.
Was the implementation process straightforward?
Our in-house development team handled the integration, and we were live in just under a month, which was much faster than we anticipated.
Results: Meaningful fraud reduction without sacrificing the customer experience
What has changed since go-live?
We launched in August, which is peak summer season for Asoview and one of our highest-volume periods of the year. Despite that timing, we still managed to achieve a significant reduction in fraudulent transaction rates within the first month of operation. Sustaining that performance during peak demand was a meaningful proof point.
What we're most satisfied with on the precision side is that false positives did not increase. We're now blocking fraud with surgical accuracy, using granular parameters like operating system, screen size, and behavioral signals without disrupting legitimate customers' purchasing experience.
As fraud-related contacts have declined, our customer support team has been able to redirect its capacity toward core service issues.
How does the operational model with DGBT work in practice?
We don't need to manage configuration changes internally. We pass fraud data and feedback to DGBT, and they handle the tuning, which has been genuinely valuable. In practice, patterns and tactics identified through daily operations are fed back into the system configuration, and the AI model learns from that input to improve detection accuracy over time. Where manual intervention is needed - say, when we need to block a very specific fraud pattern immediately - DGBT's consulting team translates that into rule logic directly.
We started with a primarily rules-based setup at launch, and as data has accumulated, the AI model's accuracy has progressively improved. Running that cycle in partnership with DGBT means we're continuously raising detection quality while keeping our own operational overhead low. Going forward, we intend to lean further into the AI model and shift toward even higher-precision detection.
A security strategy built for global scale
What does the road ahead look like?
Asoview is expanding horizontally from activities and leisure into events and travel-adjacent categories, and we're accelerating internationally. As the market grows, so does the attack surface. AI advances are creating operational efficiencies on one side and more sophisticated attack vectors on the other. That's precisely why a partnership with a solution like Sardine, one that continuously evolves, combined with the operational expertise of a partner like DGBT, is not optional, it's essential.
Fraud prevention is a continuous effort. The adversaries adapt constantly. The fraud data we accumulate feeds into Sardine's model improvements, which in turn raises our own detection accuracy and is a compounding, mutually reinforcing relationship. That's the kind of sustained, win-win dynamic we intend to maintain.