By Chris Murphy
Chatbots could soon become every bit as essential to engaging with customers as mobile apps and websites are now. Customers want to use their preferred messaging channel—Facebook Messenger, Slack, SMS, WeChat, Alexa, Google Home, and the like—to get instant answers. And thanks to advancements in artificial intelligence, deep learning, natural language processing, and back-end integration, the technology is here for companies to finally deliver smart, bot-driven conversations.
“What we’re seeing is this huge transformation—messaging as the next browser,” says Suhas Uliyar, vice president of mobile, bot, and AI strategy and product management at Oracle. “We saw the internet wave, then we saw the mobile wave, and what we’re now seeing is the messaging wave—the conversational UI, backed by artificial intelligence.”
Headquarters: Chicago, Illinois
Oracle products: Oracle Mobile Cloud Enterprise
With the intelligent bot-building capability now part of Oracle Mobile Cloud Enterprise, a multichannel development, operations, and analysis platform, companies can manage their bot, mobile, and web customer experiences as one strategy. Developers can use all their existing mobile-optimized APIs built for smartphone apps to integrate their chatbots with enterprise systems—such as building a chatbot that redirects to a mobile app to capture structured data when, for example, a chat about loan rates turns to filling out the form for a loan application. (See “Thinking About Chatbots? Keep These Five Factors in Mind.”)
And businesses eager to improve connections with customers and employees are using Oracle Mobile Cloud Enterprise to prototype and test chatbots for everything from responding to customer billing questions to providing employees with precise maintenance and status information.
Not only does [a channel-agnostic approach] allow us to deliver solutions for emerging channels quickly, but it also ensures that our customers have a consistent experience however they choose to interact with us.”–Michael Menendez, Vice President, Information Technology, Exelon
Exelon Adds Chatbots to “Channel Agnostic” Approach
A prime example of early chatbot adoption is Exelon, the United States’ largest utility by customer count, working across the energy industry in generation, electric and gas delivery, and transmission. With six utilities including Atlantic City Electric, BGE, ComEd, Delmarva, PECO, and Pepco, Exelon delivers energy to about 10 million customers across the East Coast and in Illinois.
Exelon has built a working prototype of a chatbot that relies on natural language processing and artificial intelligence to understand conversations, with integration to back-end systems for the data needed to provide clear answers. Exelon expects to let customers use a variety of messaging platforms and digital assistants to ask questions.
Michael Menendez, vice president of IT for BGE and Exelon Utilities Customer, describes the company’s technology architecture as increasingly “channel agnostic,” in which his teams can build once and quickly apply the capability across multiple channels, including new chat platforms and mobile and web apps. “Not only does this allow us to deliver solutions for emerging channels quickly, but it also ensures that our customers have a consistent experience however they choose to interact with us,” Menendez says.
Exelon built its chatbot prototype in less than two weeks using the bot-building capability in Oracle Mobile Cloud Enterprise. By using the Oracle platform, Exelon was able to reuse the microservices developed for its mobile app and use the same APIs to securely provision the needed back-end services for this new channel. “When you develop a bot application, you need to have the microservices to feed it,” says Rajesh Kumar Thakur, Exelon’s principal architect for the chatbot project.
Integration is key to the chatbot being able to offer accurate answers. Exelon relies on numerous long-running, proprietary systems for billing and outage monitoring and reporting. Oracle Mobile Cloud Enterprise lets Exelon deliver data from those systems via microservices to its customer-facing channels. The company used the bot builder’s dialog engine to craft the scripts for questions people might ask, and the machine learning capabilities help the bot refine responses over time, says Thakur.
It’s not enough just to build a chatbot, though. Having good analytics is just as essential as having a good dialog engine, because those analytics will tell marketing, operations, and customer service leaders what’s working and what isn’t with the chatbot experience.
Thakur says that access to analytics is important not only for improving the customer experience of a chatbot but also for proving the business value. “You can launch these things, but then you need to tell your business which channel is most popular or how this channel is reducing the call volume to your call center,” he says.
Oracle Mobile Cloud Enterprise includes customer experience analytics to help teams get the kind of deep insight needed to personalize engagement with end users. You want to be able to answer these very fundamental questions, Oracle’s Uliyar says: “Who is using my channel? How many users are coming through which channel? What are they doing in each channel?” And, he notes, that analysis needs to cross channels, so you know what people prefer to do on a mobile app versus a bot.
When you develop a bot application, you need to have the microservices to feed it.”–Rajesh Kumar Thakur, Principal Architect, Exelon
Calling on Instant Apps
Don’t expect to keep neat and tidy lines separating these bot, mobile, and web channels, however. In fact, Uliyar expects chatbots to accelerate a change in how mobile apps are built. Instead of a single mobile app, he talks about “instant apps” that can surface only one piece of an app’s capability. So, if you’re asking a chatbot about concert ticket options, and you pick one to buy, the bot can connect you to only that buying piece of the app, with your selection filled in, and not the full app with a seat selection map and a calendar of upcoming events.
Instant apps is an appropriate name for this idea, because the true driving force for chatbots circles back to the customer and employee desire for instant response. People text their friends because they want an answer right now, and they increasingly expect the same from companies.
Yet companies can’t afford to hire enough people to provide such instant answers. And that’s why demand is rising for this conversational UI that uses natural language and AI integrated with a back-end data system.
“I’m not advocating that human agents would go away by any means,” Uliyar says. “There’s a very good synergy between the bot and the human agent on the back end.” But thanks to advances in AI, machine learning, and natural language processing, chatbots have become a practical solution to answering more of that first wave of questions.
LEARN more about Oracle Mobile Cloud Enterprise.