theCut Slashes Fraud by Integrating Predictive AI Into Their Platform

The Old Approach

theCut had to manually analyze all transaction data to build their own custom algorithm to detect fraud. This complicated process meant that fraudulent transactions were not always detected accurately, causing the support team resources to be wasted and potentially losing more revenue to fraud.

The AI-Optimized Approach

Akkio enables theCut to start using AI in their fraud detection services, without the need for technical data science resources. Now, their fraud detection model saves time with a seamless deployment process and accurately predicts which transactions may be fraudulent, increasing their customers’ trust, saving time, and boosting revenue. 

• 30% reduction in false-positives 
• 2 months of development time saved 
• 90%+ accuracy of fraud detection model

The Challenge

theCut is a service platform that empowers barbers and shops with tools to manage and grow their business. The platform also makes it easy for users to find new barbers, schedule appointments, and pay for services in the same application. They’re ahead of the curve when it comes to technology, bringing a new narrative to the traditional men’s grooming industry. 

Any platform that processes payments is subject to fraud, and this threat is on the rise, with an estimated $48 billion in global fraud in 2023, 42% of it coming from North America. At theCut, they originally addressed this issue by using if/then statements to try to categorize and score users as fraudulent. For example, if a user created an account and 5 seconds later they booked an appointment they would be flagged for review. They then had to have their support team go and manually check all these accounts. This process without the aid of machine learning was slow and generated a lot of false positives, meaning support was not optimizing their time.

Creating the perfect fraud detection model is a delicate balance. With millions of customers and growing, theCut knew they needed a simple solution that allowed them to build and train a machine learning model based on transaction data and deploy it directly into their platform.

The Solution

When Kaushal, a Software Engineer at theCut, came across Akkio, the potential added value of AI was apparent from the start: 

“What drew us to Akkio was the ease of use. I uploaded specific data points, built my model, and within minutes it identified how to reduce the amount of false positives. We were so impressed, we had to use this tool as a part of our fraud detection process.”

After theCut connected their data to Akkio, they were able to create an AI model to predict the likelihood of fraud for transactions. The model estimates the likelihood of fraud based on the historical outcome of a variety attributes. Akkio’s model ingested data from their connected Snowflake instance such as the number of client transactions, logins in the last hour, and transaction disputes to generate the predictive model. Using these data points, Akkio’s AI automatically surfaced key insights and continuously identified patterns over time. The model is deployed via API and integrated into theCut’s booking platform, so they can detect fraud for each transaction in seconds. With the model integrated into their platform, theCut no longer had to manually analyze data and stay on top of trends, and could more accurately predict fraudulent users for the first time. 

Thanks to Akkio, theCut no longer needs to rely on manual analysis and modeling. Instead of building their own machine learning/AI solution, they can easily connect their Snowflake data warehouse, and Akkio’s AI creates an accurate model from their data that requires minimal maintenance. As a result, both the development and support teams can spend less time assessing fraud within their user base. 

“The Akkio team is quick, responsive, and customer-focused. We wanted an easy way to get our data from Snowflake to Akkio. Within a few days, Akkio developed a solution for us and the process was seamless.”

The Impact

Once theCut started using AI to model key outcomes, they were able to reduce false positives by 30%, so fewer non-fraudulent transactions were marked fraudulent. Before, they had to continuously analyze transaction data and constantly stay on top of new trends. With Akkio, the model can automatically identify any new trends, optimizing their fraud detection process – saving time and resources. 

With Akkio, theCut has found a way to optimize fraud detection in payment processing. By being early adopters of AI, they have positioned themselves as a formidable force within the men’s grooming industry – enhancing their competitiveness and helping them grow their customer base.

Interested in building your own fraud detection model? Follow the entire step-by-step tutorial, based on theCut’s process.