In the past year, we’ve seen major advances in mobility and biometrics, paving the way for less customer friction in digital transactions. Here are five predictions for what’s in store in 2018.

1. The Biometrics Market Will Be Disrupted

Biometrics is one of the hottest sectors in technology today, with crucial applications for payments, banking, retail, health care and government. While some predict that biometric revenues will surpass $30 billion in the global marketplace by 2022, I disagree.   

In November, Apple revealed that it plans to share facial mapping data with app developers. However, the agreement states that developers will only be able to access the data, not Apple’s mathematical representation of that data, which unlocks a device via Face ID. That information is encrypted on the user’s device and not even available to Apple employees.

Other device manufacturers are likely to follow Apple’s lead and open up access to mobile biometrics data. This will deal a big blow to third-party biometrics providers. Previously, app developers had to source software development toolkits (SDKs) and biometric libraries from these vendors. As these become available from device manufacturers, the need to source them from a third party will be severely diminished.

2. Rules Engines Will Disappear

With the amount of data following Moore’s Law and seemingly doubling every two years, the number of situations, rules and exceptions contained in a company’s fraud detection engine have become more than unwieldy — it’s now a liability. For example, the only way for a bank to stay ahead of fraudsters is to change a rule before it’s bypassed. In order to do this, vulnerable rules must be detected, which is virtually impossible when there are thousands of rules, many of them conflicting or irrelevant.

While machine learning engines and artificial intelligence systems can be trained to learn patterns and outliers of fraud and dynamically react to them, getting there can be daunting. It often involves ripping out legacy technology stacks built with point solutions and data providers on top of rules engines. Despite the obstacles, machine learning will displace human-defined legacy rules-based engines, whose days are numbered.