ClearSale is a company that understands that protecting business against fraud-related chargebacks takes more than what technology can provide. Earlier this month, Experian agreed to buy ClearSale for $350 million. Pulse 2.0 interviewed ClearSale’s Senior Director of Risk and Customer Success at ClearSale, Bruno Farinelli.
Bruno Farinelli’s Background
What is Bruno Farinelli’s background? Farinelli said:
“I graduated with a degree in statistics from UNICAMP, Universidade Estadual de Campinas, a very renowned university in LATAM, and went on to receive an MBA in Business Intelligence from FIAP. I was then invited to join the Analytics and Data Science team at ClearSale back in 2012. Over the past 12 years, I’ve held various positions with the company that have guided my career and personal growth. My time with ClearSale has been very rewarding and I’ve enjoyed enhancing my knowledge in fraud prevention strategies, emerging technologies, and what the future has in store for e-commerce.”
“As the Senior Director of Risk and Customer Success, I spearhead risk management and account oversight for ClearSale’s international client portfolio. To deliver exceptional customer experiences, my team and I have devised strategies that propelled record-high transaction approval rates and automated decision-making while maintaining very low chargeback levels. Our approach ensures clients benefit from seamless interactions and optimal outcomes.”
Core Products
What are the company’s core products and features? Farinelli explained:
“ClearSale is an anti-fraud technology company that provides fraud prevention and management solutions for e-commerce SMB and Enterprise businesses. Its core products and features include ML and AI-based fraud detection, manual review tools, chargeback prevention, and business intelligence reports to help retailers identify and mitigate fraud risks.”
Evolution Of ClearSale’s Technology
How has the company’s technology evolved since launching? Farinelli noted:
“When ClearSale started, our fraud detection capabilities were primarily based on logistic regression algorithm paired with rules and manually defined scenarios to identify potentially risky orders. While effective then, this rules-based approach had limitations in keeping up with constantly evolving fraud tactics.”
“Over the years, we’ve invested significantly in ML and AI tech to drive more advanced fraud detection models. Our technology now utilizes learning techniques that can analyze hundreds of data points across an order in real time to identify even the slightest anomalies or red flags that may indicate fraud.”
“Importantly, these AI models are not static – they continuously learn and adapt as new fraud patterns emerge. The models get smarter over time by ingesting insights from our global retailer network and expert analysts who closely and continuously review complex fraud cases, feeding that information into our systems that can learn and weed out that type of scheme in the future.”
Challenges Faced
What challenges have Farinelli and the team faced in building the company?
“We are constantly facing challenges from increasingly sophisticated fraudsters leveraging new technologies like AI and machine learning. Fraudsters are using AI to automate and scale up their tactics, making fraud attempts harder to detect through traditional rules-based systems that we used to rely on.”
“ClearSale has been using AI for years to identify potential fraud threats. However, as this technology has been more widely adopted, blending human expertise to analyze flagged orders has been crucial to approving as many authentic orders as possible. Ultimately, fighting AI-powered fraud requires an AI-first approach but also the flexibility to continuously evolve detections as fraudsters change tactics. It’s an arms race, but one we are determined to stay ahead of through calculated investments in talent and technology.”
Pitfalls Of Using AI For Fraud Detection
What are the pitfalls of using AI for fraud detection? Farinelli replied:
“Quality issues within the training data can lead the AI to learn and amplify biases or blind spots. There will always be a subset of ambiguous edge cases that require human judgment to properly evaluate contextual nuances and confusing customer behaviors. Perhaps most critically, fraud tactics continuously evolve, and state-of-the-art AI may lag in adapting to new threats as fraudsters leverage it themselves for automated attacks using deepfakes, synthetic identities, and more. As a result, overreliance on AI alone is risky – a hybrid model blending advanced technology with specialized human oversight and expertise is crucial for maximizing fraud prevention while avoiding costly false positives or missed threats.”
AI For Transforming E-Commerce For Retailers
How is AI transforming e-commerce for retailers and being harnessed for good? Farinelli answered:
“AI enables crafting those exceptional personalized shopping experiences customers increasingly expect. By analyzing massive datasets on browsing activity, purchases, and preferences, AI allows us to truly understand each shopper’s needs and tastes. Retailers can then easily deliver hyper-relevant product recommendations, messaging, and offers tailored specifically for each individual shopper. It creates a sense of intimacy at a scale that traditional marketing tactics cannot match.”
“AI also provides immense value in fraud prevention as it can sift through hundreds of data points in real-time to aid in our efforts to fight against fraud. It’s about using advanced technologies to deliver tailored convenience securely, not replacing human elements entirely.”
Differentiation From The Competition
What differentiates the company from its competition? Farinelli concluded:
“A key differentiator is our proprietary advanced review process for high-risk transactions that is overseen by a global team of highly trained fraud analysts. While many fraud prevention companies simply decline the majority of orders flagged by their systems as potential fraud, we take a nuanced approach–leading-edge technology and expert review processes that ensure no good order is declined.”
“This hybrid model allows us to minimize the number of valid orders being rejected, reducing false positives that can harm customer relationships. Our advanced secondary review process with expert analysis makes precise decisions to ensure the right orders get approved and the fraudulent ones are canceled before harming the retailers.”
