Direct video surveillance of shoppers has long been the holy grail of fraud prevention, and today–after years in development–Riskified is thrilled to unveil the latest advent in fraud-prevention technology: ‘Cameo’. Cameo is a proprietary fraud detection technique powered by behavioral analysis of online shoppers, which can automatically determine the legitimacy of an order with extreme accuracy.

Early trials of Cameo have yielded exciting results, and we expect this surveillance data will allow us to dramatically improve our merchants’ approval rates while nearly eliminating CNP fraud.

How does it work?

Riskified has partnered with a well-known Russian security consulting firm to develop covert surveillance techniques, which allow us to gather enormous amounts of behavioral data on shoppers. The technology utilizes pre-existing cameras and microphones embedded in laptops, phones, CCTVs, and more, to capture footage of online shoppers’ movements, facial expressions, and environment. By working with body language experts, military interrogators and professional poker players, we’ve developed a robust database of legitimate and fraudulent behaviors, which we use to enrich our more traditional forms of data.

New insights into fraudster behavior

Thanks to thousands of hours of footage provided by Cameo, we’ve observed several surprising trends of legitimate and fraudulent shopping behavior:


  • Fraudsters lick their lips much more frequently than legitimate shoppers. Additionally, on average they blink twelve times per minute, compared to ten for legitimate shoppers.
  • Pupil dilation and flushed cheeks are indicators of risk, likely due to the thrill of crime. However, we also frequently observe these behaviors in legitimate shoppers of Apple products.
  • Heavy metal blaring in the background raises the risk of a fraud attack by 23%. Conversely, smooth jazz is a sign of legitimacy. Orders placed while shoppers are humming along to Kenny G can be approved 99.4% of the time.
  • Both fraudsters and legit customers are equally likely to be eating spaghettios directly out of the can while shopping online.

 The future of fraud prevention

The implications of a fraud prevention solution that can surveill users are endless. Not only can we observe while customers are engaged in the shopping process, we can watch during most hours of the day. The advent of big data and machine learning are enabling us to know the likelihood of someone committing CNP fraud hours, or even days, beforehand – sometimes before the idea has even occurred to them.

Riskified’s Cameo-enhanced fraud solution will deliver increasingly firm messages to would-be fraudsters, first inquiring gently if they’re sure this is a road they want to go down, until eventually informing the would-be fraudster that sorry, but we’re afraid our algorithm simply can’t allow them to do this.

Ethical concerns


Get started with Cameo today!

If you’re a merchant interested in requesting a demo of our Cameo-powered fraud prevention solution, we already know, thanks to big data, machine learning, and dozens of micro cameras installed in your household appliances. We’ll be getting in touch shortly.