ServiceNow launches machine learning, AI automation engine

The company is using technology from its DxContinuum acquisition to build an Intelligent Automation Engine that tailors models to each customer.

Today’s State of Work: At the Breaking Point

ServiceNow is rolling out a machine learning engine that's designed to predict outages, automate routing and workflow, predict outcomes, and benchmark performance.

The engine, called the Intelligent Automation Engine, was built with technology from ServiceNow's DxContinuum acquisition in January.

ServiceNow said that it will use the Intelligent Automation Engine across the platform and in cloud services for customer service, security, and human resources as well as IT.

While many machine learning and artificial intelligence efforts have been set up as add-ons to cloud services, ServiceNow CTO Allan Leinwand said the Intelligent Automation Engine will be tailored for each customer account.

"We're not building a piece of AI that everyone shares and we're not using a data lake. We're taking individual data out of customer instances, shipping it to a training engine, building a model and putting it into that instance," said Leinwand. "Once the model is in the instance it is in memory."

The Intelligent Automation Engine arrives as automation is becoming a front-and-center issue for companies. According to a ServiceNow survey, 86 percent of companies say they will need more automation to get work done by 2020. Why? The volume of work will need machine help.

In ServiceNow's survey of 1,874 global executives it found that half say the pace of work has increased 20 percent in the last year, but only 42 percent of business processes are automated on average. As a result, leaders are spending up to 16 hours on manual tasks.


The four main use cases for ServiceNow's automation efforts include:

  • Anomaly detection to prevent outages in IT departments. ServiceNow will apply algorithms to find patterns and outliers that can lead to an outage. Anomalies can also be correlated with past events and workflows.
  • Routing and categorizing of work. Learning algorithms will automatically route work based on past patterns. Tasks such as assessing risk, assigning owners, and categorizing work will be automated.
  • Performance predictions. The Intelligent Automation Engine can be used to set a performance goal and data profile and get predictive analytics on hitting goals.
  • Benchmarks vs. peers. ServiceNow is using the automation engine to compare companies to their industries and peers to gauge efficiency and make recommendations.