Workday Rising previews AI tech that draws from customer data sets
Workday is developing algorithms to automate certain tasks, such as anomaly detection. This technology was previewed at Workday Rising, with the tech planned to arrive in 2019.
LAS VEGAS -- Workday is making machine learning and pattern recognition key capabilities in its platform. The algorithms will be trained using customer data to spot changes, such as anomalies in key metrics, and alert users to problems. This technology was a major theme at the vendor's Workday Rising conference last week.
Business analysts today run reports on many of the same things that the forthcoming algorithmic systems can accomplish. The difference is this AI-related technology will run the queries automatically and bring changes to a company's attention that deserve a closer look.
The technology will become available gradually to Workday customers. One announced product, Workday People Analytics, may not be available until 2019. Workday will begin testing it with early adopters this spring.
Not revolutionary, but a good step
This technology represents a potential shift in how HR and finance users work, but how much of a shift remains to be seen.
"It's not going to revolutionize finance, but it will certainly be a good step in the direction," said Daniel Berringer, controller at the Association of American Medical Colleges in Washington, D.C., after one Workday Rising session devoted to the finance platform and the recent Adaptive Insights acquisition.
Berringer said the tech may be able to pinpoint certain irregular transactions before a financial close, and that alert could be helpful.
At the conference, Workday also announced it is coupling finance and HR data analysis as a result of acquiring Adaptive Insights, a vendor of corporate performance management software.
To get to this point, Workday isn't advancing AI-related technology. Instead, it is taking advanced technology available in the industry and training the algorithms to do specific jobs. This data is its advantage, Workday officials said. It has more than 2,300 customers working on standardized data models, and that gives Workday access to what it needs to train machine learning and pattern recognition algorithms.
"We're not inventing fundamentally new techniques in machine learning or AI," David Clarke, senior vice president of technology and infrastructure at Workday, based in Pleasanton, Calif., told reporters at Workday Rising. "What we have that's unique are the data sets that we have access to," he said.
Workday Rising attendee likes potential of tech
What we have that's unique are the data sets.
David Clarkesenior vice president of technology and infrastructure, Workday
Workday users will get a dashboard that has reports on any number of metrics. Take, for instance, this HR category: regrettable attrition of new hires. This term refers to new employees a firm wants to retain, but who leave after a short period.
If regrettable attrition data changes in ways that raise flags, the Workday platform will bring it to the attention of users. It will do detective work, as well. The system will have numerous other data points, including geography, gender, the supervisory organization and other categories, that may help explain why new hires are quitting. The same type of approach works in finance.
The intent of the system is to produce details "that people can act on right away," said Pete Schlampp, vice president of Workday Analytics, in an interview.
Along with pattern recognition and machine learning, Workday is using graph processing technology, which creates a massive graph showing how all the data points connect. A data point might be that an employee didn't like the manager or the compensation -- these are nodes on a graph, Schlampp said. The system uses natural language processing to turn these discoveries into plain-language explanations.
One user at the conference, Brent Park, director of finance operations at United Allergy Services, based in San Antonio, said the technology may be "incredibly useful." He cited potential uses, such as identifying changes in the cadence of customer payments, and he said this tech is "taking anomalies to a trend."
Workday People Analytics will be sold as a stand-alone product to HR customers, but it will be bundled with Workday Prism Analytics, an existing platform. There will be limited availability next fall, with general availability to follow, Workday officials said.
