TORONTO — Waterloo, Ont.-based OpenText kicked off its EnterpriseWorld 2017 conference here with the long-awaited launch of Magellan, its platform for introducing artificial intelligence- and machine learning-based analytics to the business data residing on its platform.
Magellan works on a simple equation, said OpenText CEO Mark Barranechea. “Information plus algorithms equals insight,” he repeated several times during his Tuesday morning presentation. The idea, he said, is to make enterprise data lakes easier to “fish” and “explore,” bringing deeper analytics — and particularly predictive analytics — to everything customers store in the company’s enterprise content and enterprise information management systems.
“You take your enterprise information management information, you apply an algorithm, and you get insight,” Barrenechea said.
Algorithms can be programmed in a variety of languages, including Python and R, making it relatively easy to find the needed/wanted information from the lake pulled together from OpenText’s content management systems.
And it’s in that programmability — indeed, Magellan by its very nature will require skills to both connect the system to the right data, and then pose the right questions to that data set in terms of algorithms and applications — where the company’s partners will prove invaluable.
Barrenechea said Magellan should be “a boom” for the company’s partner community, offering “a whole new platform to bring” to their area of specialization. While OpenText will no doubt do some professional services and development work on bringing Magellan-derived applications to certain key industries, Barranechea said the company’s partners will be key in taking the message broader.
“I would hope that a thousand applications bloom, and 10,000 algorithms bloom” out of Magellan, he said. “Magellan is amazingly partner-friendly, and opens up a whole new platform for partners to bring to their customers, a whole new aspect to the relationship” with their customers.
The company has firmly targeted rival IBM and its Watson cognitive computing platform for the announcement, with Barrenechea pointing out several points of differentiation between Magellan and Watson, the most important one of which is “you don’t need to hand over your intellectual property and algorithms to IBM,” as Magellan can run in a variety of modes, ranging from on-premise to fully OpenText cloud-based, depending on customer preferences and requirements.
The infrastructure required is another key differentiator, as Magellan is designed to run on x86 servers, with a number of pre-certified configurations from various vendors.
“You don’t need a mainframe or a proprietary cloud, so incased of spending $1 million in hardware, this is about $100,000,” he said.
Identifying opportunities should not be too hard — Barrenechea said anywhere customers are doing predictive analytics with spreadsheets, there’s an opportunity.
“We think our customers know their data and the algorithms they want to apply,” he said.
One such early customer is – surprise – the company itself. CMO Adam Howatson granted that he got “a special deal” on licenses for the software, but said he saw a lot of value in using Magellan to fine-tune its own marketing engine. First up, he said, he wants to let Magellan loose on every RFP the company has received in its 26-year history, to boil down what it can do better moving forward.