OpenText launches Magellan, an AI platform aimed at IBM's Watson

OpenText is betting open source, Apache Spark, and years of experience with unstructured enterprise data will enable it to take on IBM's Watson and land midmarket customers.
Written by Larry Dignan, Contributor

OpenText is launching an artificial intelligence platform called Magellan that hopes to use the company's knowhow with unstructured and semi-structured data and open source to compete with the likes of IBM Watson.

At EnterpriseWorld, OpenText's conference in Toronto, the company outlined its artificial intelligence and machine learning platform.

OpenText focuses on enterprise information management and has a portfolio that extends into content management for industries, customer experience, and data discovery.

In January, OpenText completed the purchase of Dell-EMC's enterprise content unit, which includes Documentum.

For the third quarter, Open Text reported revenue of $593 million, up 35 percent from a year ago. OpenText CEO and CTO Mark Barrenechea said in a recent interview that the company has filled in functional gaps with acquisitions and sees information management as a key pillar for digital transformation projects and automation.

Artificial intelligence is also an extension to OpenText's specialty. "We've avoided fashion statements in technology, but we've embraced some early. AI is very real and practical," he said.

OpenText is using an open source approach with Magellan with integration with Apache Spark and MLlib, a machine learning library. "We are combining the strengths of OpenText and the open source community," said Adam Howatson, chief marketing officer at OpenText.

Magellan's approach will be to enable customers to leverage open source intellectual property and algorithms as well enabling companies to build their own models. Howatson added that OpenText's Magellan platform will have a lower price point, be available as an appliance and be available on premises or via the cloud.

Lower costs associated with Magellan are likely to appeal to large companies as well as the mid-market, a segment of companies that aren't likely to have a dozen of data scientists.

The differentiator for OpenText will be the ability to deploy text mining and analysis to unstructured and semi-structured data. With Apache Spark and open source as a backbone, OpenText customers will save on training and integration. "We don't need to start from scratch," said Howatson. "Since it's Apache, we don't need a bus load of integrators."

Magellan will also include technology from Nstein, an acquisition that brought text mining and auto classification to OpenText, and Actuate, an analytics platform. OpenText has been built via acquisition.


"We've captured massive amounts of data and now AI and machine learning platforms can complete the vision of enterprise information management," said Howatson.

Use cases abound, but one obvious one for Magellan will be to analyze requests for proposals. A chief marketing officer can use AI to take the unstructured data in an RFP and understand what customers want. "A human can't get there because there's too much information," said Howatson.

Editorial standards