Subscribe

Share

Emerging Technologies

Thinking About Chatbots?

By Chris Murphy

November/December 2017

IT and business-unit leaders thinking about the platform they’ll need to develop, run, and analyze chatbots should consider the following five factors, according to Suhas Uliyar, vice president of mobile, bot, and AI strategy and product management at Oracle.

  • Dealing with multiple channels. Your tool should let you write a chatbot once that can run on many channels, such as WeChat, WhatsApp, and Facebook Messenger, as well as voice assistants such as those powered by Alexa and Siri.
  • Having real conversations. Developers need a dialog engine that uses natural language processing to help you write intelligent scripts at the start—that can understand and respond in conversational language.
  • Using machine learning. Your dialog engine also needs algorithms that can learn the many different ways customers ask questions, so it gets smarter over time.
  • Integrating with enterprise systems. A bot’s intelligence comes from the data in systems such as customer relationship management, supply chain, enterprise resource planning, and more. Your bot needs to tap into a flexible, API-driven integration platform.
  • Enabling bot-to-human-agent handoffs. Your bot should be intelligent enough to know when it can’t handle an end user’s input and then hand off to a human agent, to ensure the best-quality experience. And the human agent should be able to hand off conversations to the bot when appropriate.

“At Oracle, we focus on making it very simple for developers to build bots, instead of having to build all these five things from the ground up. That’s what we have built into our chatbot platform,” Uliyar says. “I think the secret sauce of Oracle Mobile Cloud Enterprise is the sum of all its parts. The differentiator for the platform really is a packaging of all this into one cohesive solution.”

Next Steps

LEARN more about Oracle Mobile Cloud Enterprise.

Photography by Edwin Remsberg/The Verbatim Agency