Oracle puts AI center stage at OpenWorld

If you’re an Oracle customer, you’ve probably been invited to move some of your spending with them to a subscription model. Now Oracle wants to make it easier for you to move your own customers to subscription-based pricing too.

Oracle Subscription Management is a new cloud application the company unveiled at its OpenWorld event in San Francisco this week. Oracle also upgraded existing cloud apps with a host of new automation and personalization features based on artificial intelligence, and introduced new blockchain-based supply chain management tools for track-and-trace and lot provenance.

Subscribe to everything

Subscription-based business models encompass everything from car insurance to cars themselves these days, and soon automakers may even buy subscriptions to the machines in their factories, thanks to the industrial internet of things. But they can pose problems for enterprises that have built their businesses on software designed for one-off sales.

“We saw more of our customers moving from product-centric companies to subscription companies,” said Steve Miranda, Oracle’s executive vice president for applications and product development. “We have a brand-new subscription management system to handle orders, revenue recognition, payment, everything for the full life-cycle of subscriptions and contacts.”

The new cloud app links back- and front-office systems to provide customers and customer-service staff with access to subscription management features. That, Oracle hopes, will make it easier for sales staff to see what customers are already buying, and propose a subscription offer instead. From within Oracle’s core cloud ERP system, business managers will be able to access analytics on subscription performance, and recognize subscription revenue in their accounts.

Adding AI to ERP, SCM and recruitment

As for Oracle’s use of AI in its cloud applications, “It’s a model-driven approach where the application starts to learn based on what you do, the context of what you do, and personalizes that towards your use,” Miranda said in his keynote presentation at OpenWorld. “We believe that this is going to fundamentally change the way you set up, configure and change the applications from purely an input-and-validate application to a push-and-suggest application.”

AI-based functionality is being increasingly woven into ERP solutions. Among the ERP functions Oracle is automating using machine learning are process automation, payment processing and access control management. Supply chain management is another prime target for AI features, and Oracle is now offering automation tools around supplier recommendation. In HR the company is using machine learning to identify candidates with the best fit for a job based on criteria that go beyond simple keyword analysis.

Chris Pang, senior research director at Gartner, sees real business benefits here. “Oracle is trying to take [CV analysis] beyond skills matching and look at it in terms of experience,” he said. The recruiting teams in many organizations can’t keep up with requirements, he said, and a tool like this will help reduce the time to fill vacancies.

Oracle expects the software to learn over time which recruits were successful and to look for more like them.

Letting machines loose in areas such as this can be risky: “You have to be careful you don’t introduce bias,” Pang warned.

Smart accounting

Oracle is also using machine learning to adapt user interfaces according to time and context, and to build bots that can connect to conversational interfaces, including text messaging and voice interfaces.

One bot that Oracle CTO Larry Ellison himself will demonstrate later this week is designed to help file business expenses via smartphone. From within a conversational interface such as Slack, it uses the phone’s camera to scan a receipt, and recognize the date, amount, vendor and purpose of the expense. It then attempts to categorize the expense. If it recognizes a meal, for example, it can classify it as lunch or dinner based on the timestamp and enter the number of attendees if that is shown.

“Then it uses machine learning to enforce policy rules based on your own organization’s approval system,” Miranda said. “It’s going to be interesting on Wednesday to see Larry’s meal, to see if that’s within expense policy or not.”

Specialist software vendors are already using AI techniques to improve the accuracy of optical character recognition on scanned receipts, according to Gartner’s Pang.

“Oracle already had OCR before, and what they have done is put better machine learning on top of it. This is more of an evolution than a revolution,” he said.

For enterprises, though, it represents an opportunity to simplify things by staying within Oracle’s software suite rather than going best of breed, he said.

More broadly, Oracle is using machine learning to identify payments in other business processes that need additional approval, or to identify transactions that users usually approve and ask them if they would like to automate that step in future.

Of course, Oracle is not alone in adding AI functionality like this. SAP’s Concur division began beta-testing an expenses bot operating in a Slack channel back in April, and earlier this month QuickBooks said it will use machine learning to power similar transaction processing systems for users of its ERP system for small businesses. Also this week, SAP said it will soon offer a conversational AI for the enterprise that can be used to build customer service bots tied to the heart of its ERP system.

Track and trace on the blockchain

Back in July, Oracle opened its platform-as-a-service Oracle Blockchain Cloud Service for business. Now it’s making it simpler for enterprises to adopt it with the release of four ready-made services that integrate with its ERP and SCM cloud apps: Intelligent Track and Trace, Lot Lineage and Provenance, Intelligent Cold Chain and Warranty Usage & Tracking.

One of the first users of the track-and-trace app was a treat for OpenWorld attendees: Californian brewery Alpha Acid Brewing supplied beer in QR-code-marked bottles for the OpenWorld event, with the idea that scanning the code with a smartphone would send drinkers to a web page listing where the beer’s ingredients came from.