Infor’s artificial intelligence strategy: advise, augment, and automate
July 12, 2017
Artificial intelligence (AI) is real, and it’s going to be huge.
If you think you’ve heard that one before, hold on a minute. It’s true that AI has been hyped as a game-changing technology since the 1970s, but now it’s really here, and it may even be in your home right now.
AI is evident in applications like Amazon’s Alexa and Google Assistant, which understand spoken commands nearly perfectly and filter responses to questions based upon preferences they learn over time. It’s evident in self-driving cars, which will be common on many public thoroughfares within a couple of years. And it’s evident in recent victories by computers over people at games like Jeopardy! and Go, which were once thought to be impervious to algorithmic competition.
“Sometimes things take many years to mature,” said Infor CEO Charles Phillips in Tuesday morning’s keynote General Session at Inforum 2017. “This technology is starting to creep into our lives and train people to get used to it.”
It’s creeping into Infor’s technology stack as well. The company this morning announced Coleman, an enterprise-grade AI platform that applies machine-learning technology to help Infor’s customers optimize and streamline their businesses.
Named after Katherine Coleman Johnson, the physicist and mathematician whose contributions to the early NASA space missions were featured in the Oscar-nominated movie Hidden Figures, Coleman is designed for the industrial and vertical-market challenges that are Infor’s core competency. It uses image recognition, voice interaction, and automation to respond to questions, recommend improvements, and take full control of some basic tasks at the user’s option.
“Infor Coleman is AI that maximizes human potential,” Phillips said.
Coleman is designed to augment human decision-making in four areas: automation, augmentation, interaction, and advice, said Infor President Duncan Angrove. Machine learning is a true game-changer because it removes the need for humans to tell the computer what to learn.
“Traditionally, we taught machines to do things by writing programs, which was painful,” he said. “Machine learning is about teaching machines to teach themselves. The software learns from the data, accumulates knowledge and gets smarter.”
The approach is similar to trial and error, but with a twist. Machine-learning algorithms repeatedly sort through large volumes of data to look for patterns that might be meaningful. They discard useless associations and iterate on promising ones. As the machine “learns,” additional data can be introduced to refine its knowledge base, or crowd-sourcing can be used to validate good data. That’s how machines can learn to translate languages without much human intervention by conducting probabilistic analysis on increasingly large bodies of data fed to them over time.
When applied to business tasks, Coleman will dramatically improve productivity, Angrove said. For example, “It can observe hundreds of millions of invoices being matched and learn from those observations what the optimal matches are for the organization,” he said. “It can reconcile journals, expense reports, deductions, and even help close the books. It frees people from low-value repetitive tasks.” Coleman will also be able to respond to voice requests for information, reducing the 8 to 10 hours that McKinsey has estimated office workers spend looking for information each week.
For Infor customers, the biggest payoff may be the integration of Coleman into the company’s massive store of data from across industries. “It learns from our petabytes of industry information using the cloud as a supercomputer,” Angrove said. “This will bring recommendation and optimization power to each area of the business.”
Infor’s cross-industry expertise will thus help customers benefit from the experiences of other customers in completely different industries. For example, it can optimize asset deployment schedules based upon successful demand prediction paradigms from other industries. “We believe it will shatter traditional time-series-based approaches to optimization,” Angrove said.
Some elements of Coleman are available today, such as predictive inventory management for healthcare, price optimization management for hospitality, and forecasting, assortment planning and promotion management for retailers. Coleman will be integrated into Infor’s CloudSuite industry suites over the next year.