Amid the fierce competition in the identity management market, OneLogin Inc. is turning to artificial intelligence for an edge.

The company unveiled a new iteration of its authentication service today that employs machine learning algorithms to fine-tune how much scrutiny is applied to users when they try to sign into a service. According to OneLogin, the technology works by factoring the locations from which a worker typically accesses an application, what devices they use and various related information into a profile of their activity patterns. Each new login attempt is then compared against the model to determine whether it’s legitimate.

A similar approached is employed by the recently funded Fortscale Ltd. and other breach prevention startups to detect malicious activity in corporate networks. The main difference is that OneLogin places a specific focus on convenience. When users try to access a service in a way that fits with their usual habits, they only have to input anything their identifier and password, while login attempts made under different circumstances will require additional verification.

The number and type of identity checks that OneLogin performs depends on how much risk its algorithms perceive. If, for example, a worker who normally uses an application during business hours tries to access it at an unusual time, the startup’s service may merely ask the worker to answer a security question. More suspicious sign-in attempts in turn can lead OneLogin to require another form or proof such as a one-time password sent to the user’s handset or a physical authentication device.

Organizations can customize the process according to their security requirements. OneLogin provides the ability to block logins that originate countries where a firm doesn’t have business operations. The startup is looking to kill two birds with one stone and help companies increase application security while minimizing the amount of hassle that users must go through to access work-related applications.

In the same spirit, OneLogin is releasing an update for its mobile client alongside the machine learning feature that aims to speed up authentication on smartphones and tablets. Users will now receive onetime passwords sent by the startup’s service in push notifications that won’t require them to launch the app as was the case until now.

Image: Pixabay