Google Cloud made big changes to its fraud-detection system in response to an angry customer complaint that went viral

Google Cloud chief Diane Greene

caption
Google Cloud chief Diane Greene
source
Steve Jennings/Getty

  • Google’s cloud platform drew a lot of criticism after one of its customers, an admin of an account that was erroneously flagged for fraudulent activity, posted an angry post to Medium that went viral.
  • On online message boards, critics charged that the platform relied too heavily on automated systems and complained that it was too hard to reach humans during emergencies.
  • Google Cloud execs promised to review and improve its policies and they appear to have done that.
  • On Wednesday, Google announced changes to the way it responds when fraud is detected, and the changes include adding more humans to the equation.

The Google Cloud Platform (GCP) has changed the way its abuse-prevention arm responds to suspicious activity, and from now on it won’t be so quick to shoot first and ask questions later.

More importantly, Google has added more humans to its fraud-detection mix, so those overseeing accounts mistakenly accused of fraudulent activity can reach out for help, 24/7.

A lot of criticism was directed at GCP earlier this month when a customer complaint posted to Medium went viral. Thepost was writtenby an anonymous administrator overseeing a system that monitors “hundreds of wind turbines and scores of solar plants.” The admin said Google blocked the system’s website, app, and other services on June 28 without warning because it had detected “potential suspicious activity.”

The system admin wrote that Google then threatened to shut down the account for good unless the service could provide ID and other documents. Nothing like that ended up happening, however, as Google fixed the issue in less than an hour.

Human touch

Protecting the reputation of Google’s cloud service is vital as the company looks to compete with larger competitors such as Amazon Web Services and Microsoft’s Azure cloud unit. But critics of GCP noted that its customer service was often hard to reach and the company asked customers to rely too heavily on automated systems in emergency situations.

Within days of the backlash, GCP managers began a review of its procedures and promised to address the problem. True to their word, GCP execs said in a blog post on Wednesday that the company would no longer give automated systems the power to suspend accounts controlled by “established customers with offline payment.”

The company also said that accounts with an “established payment history” that appear to be engaged in fraudulent activity will receive advance notice and five days to address the problem.

“These changes are intended to provide peace of mind and a predictable, positive experience for our GCP customers while continuing to permit appropriate suspension and removal actions for the inevitable bad actors and fraud,” GCP execs wrote in the blog post.

Perhaps the most eye-catching change is that accounts flagged by fraud-detection systems will have the ability to contact chat support available round the clock. In addition, before a flagged account is suspended, Google added a second human review to the process.

In short, the GCP community asked for more human oversight and Google appears to have delivered.

GCP execs added that “We’re also modifying who has authority to suspend an account, as well as refreshing our training for the teams that review flagged accounts and determine response actions; re-evaluating the signals, sources, and the tools we use to assess potential fraudulent activity.”