BETA
This is a BETA experience. You may opt-out by clicking here
Edit Story

Oracle’s ‘Pragmatic’ Approach To AI In Business Transactions

Oracle

SAN FRANCISCO—Guiding Oracle’s development of its latest generation of cloud applications are three main business imperatives: help customers innovate rapidly, create nimble processes, and make the most of their mobile, social, and other communications channels.

Speaking at Oracle OpenWorld, Steve Miranda, executive vice president of applications development, emphasized the considerable work the company has done incorporating machine learning algorithms into its comprehensive, tightly integrated suites of cloud applications. “We’re ready to run your business in the cloud,” Miranda said.

Customers experience those machine learning capabilities in three ways:

1. An intuitive, easy-to-use, voice-enabled user interface that runs on various computing platforms but is especially suited to mobile devices.

2. “A single digital assistant with multiple skills,” such as directory assistance, expense reporting, and sales forecasting. The Oracle Digital Assistant is an interactive agent that incorporates sophisticated “bot support”—ML-enabled conversational interactions, either by voice, text, or conventional typing, Miranda said. 

3. A machine learning-based recommendation engine that interprets the Digital Assistant’s conversational interactions, taps the appropriate cloud application or applications, corrals meaningful data from both internal and external data sources, then generates suggestions—“best action,” “best fit,” or “best next step”—related to completing the transaction. It’s even able, if programmed so, to initiate actions based on those recommendations, he said.

Miranda demonstrated a series of “first-thing-in-the-morning” interactions on his cell phone. He asked the digital assistant: “What is my daily briefing?” Tapping into a business intelligence application, the system graphically displayed data exhibiting a problematic disconnect between product demand and hiring in Latin America. “What do you recommend?” Miranda asked, resulting in a graphic representation of potential hiring spots across Latin America.

Recommendation requests represent data that ML algorithms can learn from, Miranda explained. Oracle is tracking the kind and number racked up so far by its customer-facing applications: 4.4 million recommendations in sales; 624 million in service; and 500,000 in marketing, he presented in a slide.

Oracle is bucking AI trends in business in several ways. Miranda proceeded to list several artificial intelligence myths:

• Customers must build such systems themselves. By eschewing the DIY AI application programming interface approach, and instead building AI features into its applications that are constantly learning, Oracle intends to “incrementally improve” the effectiveness of its customers’ business transactions, Miranda said.

• AI is magic. The more appropriate term for what Oracle is doing with its applications is the AI subset machine learning, Miranda said. It’s an insight-oriented, continuous-learning strategy intended to “improve the way you do business in a very pragmatic, very effective way,” he said.

• AI eliminates the need for people. “Our ML helps you become more efficient, helps you realize things maybe you didn’t know before,” he said. “But ultimately you have to have people driving the innovations, driving the business decisions.”

• The more data, the better. “It’s the right data, at the right time, in the right context,” Miranda said.

ML-Imbued Applications

At Oracle OpenWorld, the company announced ML-based upgrades to a number of cloud applications:

• Digital Assistants added to Oracle ERP Cloud help reduce the effort required to submit and review timesheets, track the status of projects, and escalate time-entry and project-management issues.

Digital Assistants built into its customer-facing applications let salespeople use voice commands to create configured quotes, as well as get service agents quick answers to questions stored in the knowledge repository.

Digital Assistants added to Oracle HCM Cloud applications improve onboarding tasks, employee goal-setting, and performance evaluations. They also provide managers with self-service actions related to personnel moves, such as promotions.

“We are building AI or ML features everywhere,” Miranda told the audience, “literally every step of the way adding ML into all of your transactions to make them much, much more effective.”