Even if you haven’t held a conversation with Siri or Alexa, you’ve likely encountered a chatbot online. They often appear in a chat window that pops up with a friendly greeting:
- Thank you for visiting our site. How can I help you today?
Depending on the site, you might ask about a broken appliance in your home or inquire about investment advice. The chatbot is programmed to respond accordingly and even ask follow-up questions.
From customer service chatbots online to personal assistants in our homes, chatbots have started to permeate our lives, and they’re often helpful. But how do they work? And how can we expand their capabilities in the business world? In this article, we’ll explain the technologies that power chatbots and explore how you can use them with analytics.
Chatbots are a form of conversational AI designed to simplify human interaction with computers. Using chatbots, computers can understand and respond to human input through spoken or written language.
“Chatbots can be programmed to respond to simple keywords or prompts, or to hold complex conversations about specific topics,” says Mary Beth Moore, an AI and Language Analytics Strategist at SAS. “They range in complexity from information retrieval using keyword matches to active learning capabilities that provide in-depth responses and tailored suggestions based on previous conversations.”
Many industries use chatbots to improve or streamline customer service and e-commerce. Consider these primary applications for chatbots:
- Customer service chatbots: Many businesses are using chatbots as a first contact when customers need help. In almost every industry, companies employ chatbots to help customers easily navigate their websites, answer simple questions and direct people to the relevant points of contact.
- E-commerce chatbots: Retail companies and telecommunication providers use chatbots as an additional interaction channel with their customers. The bot is designed to lead customers through a linear process flow to complete requests or transactions.
- Virtual assistant chatbots: Personal assistants like Siri, Cortana and Alexa have risen in popularity as their benefits have become readily available and easily embedded into the daily life of consumers. People use them to quickly retrieve information, schedule appointments and interact with smart home features.
Chatbots are programmed to simulate human conversation and exhibit intelligent behavior that is equivalent to that of a human. Mary Beth Moore AI and Language Analytics Strategist SAS
How do chatbots work?
Natural language processing is a branch of artificial intelligence that teaches machines to read, analyze and interpret human language. This technology gives chatbots a baseline for understanding language structure and meaning. NLP, in essence, allows the computer to understand what you are asking and how to appropriately respond.
“Chatbots are programmed to simulate human conversation and exhibit intelligent behavior that is equivalent to that of a human,” says Moore. “With developments in deep learning and reinforcement learning, chatbots can interpret more complexities in language and improve the dynamic nature of conversation between human and machine.”
Essentially, a chatbot tries to match what you’ve asked to an intent that it understands. The more a chatbot communicates with you, the more it understands and the more it learns to communicate like you and others with similar questions. Your positive responses reinforce its answers, and then it uses those answers again.
Download: How to Do Deep Learning With SAS®
Learn more about deep learning, a key chatbot technology, and its applications. This white paper is geared towards data scientists and includes step-by-step overview of how to build a deep learning model using SAS.
Chatbot applications in analytics
Personal assistants like Siri and Alexa are a complex type of chatbot designed to respond to a wide range of scenarios and queries, from current weather and news updates to personal calendars, music selections and random questions. Chatbots with a specific purpose, like routing customer complaints or inquiries, are designed with a more limited scope of potential answers and replies.
At SAS, we’re developing different ways to incorporate chatbots into business dashboards or analytics platforms. These capabilities have the potential to expand the audience for analytics results and attract new and less technical users.
“Chatbots are a key technology that could allow people to consume analytics without realizing that’s what they’re doing,” says Oliver Schabenberger, SAS Executive Vice President, Chief Operating Officer and Chief Technology Officer. “Chatbots create a humanlike interaction that makes results accessible to all.”
Introducing chatbot functionality into analytics solutions provides a number of capabilities that marry analytics with conversational capabilities:
- The chatbot can automatically query and describe large corporate or public data sets.
- Users can request summarized or analyzed results verbally by saying, for instance, “Which marketing campaigns are generating the most leads this quarter?”
- The chatbot can provide the answer and then offer additional information or suggest a related report to view based on patterns in the data and in previous related queries.
- You can ask the chatbot to share the results with others, and that will happen automatically.
You can even combine chatbots with specialized analytics solutions to perform explicit tasks within the application.
Chatbot answers solar farm questions
How much power will the solar farm produce next month? What is the status of each solar cell? Can a solar farm generate power at night? See how a chatbot answers these and other questions in a demo from Jared Peterson, Senior Manager of Advanced Analytics R&D at SAS and Oliver Schabenberger, SAS Executive VP, COO and CTO.
Talking to solar panels
A facility manager at a large solar farm needed a tool to monitor and control solar panels when away from the computer. A chatbot was designed using SAS® for facility management that analyzes live streaming data from the solar panels.
The facility manager can now ask the chatbot directly about equipment status and each panel’s energy generation, then receive a summary of the energy output by day, month or season. A mobile chat interface extends the capabilities to technicians so they can query the application when they’re outside restarting a panel or monitoring the condition of panels in extreme weather conditions.
More time to fight fraud
One financial services nonprofit combined a chatbot with analytics to help reduce identity theft. The chatbot interacts with victims online or in phone calls to coach them through the proper protective steps, depending on their situation. It identifies cyberattacks, scams, frauds and privacy issues and the actual loss, ranging from wallets to passwords.
Based on this analysis, the chatbot asks specific questions to collect required information and recommends next steps to the victim. This streamlined approach gives humans in the call center more time to focus on case resolution.
It’s all about communication
Both of these bots use natural language processing to anticipate conversational topics and manage a typical conversational flow. They respond by giving answers or advice based on results calculated with analytics or AI algorithms.
“Chatbot technology makes AI accessible to all of us, and it uses AI to do it,” says Schabenberger.
All kinds of organizations are starting to see exciting possibilities from combining chatbots with AI analytics. But even though the technology – NLP, AI, deep learning – seems complicated, it all goes back to one simple concept: communication.
- CECL: Are US banks and credit unions ready?CECL, current expected credit loss, is an accounting standard that requires US banking institutions and credit unions to estimate life-of-loan losses at origination or purchase.
- How health care leaders deployed analytics when crisis hitDuring the COVID-19 pandemic, some health care providers were well-positioned to respond to rapid changes in demand. The factor that most distinguished them was that they already had a strong capacity in place for using data to inform decisions. Read about three key takeaways from their experiences.
- ModelOps: How to operationalize the model life cycleModelOps is where analytical models are cycled from the data science team to the IT production team in a regular cadence of deployment and updates. In the race to realizing value from AI models, it’s a winning ingredient that only a few companies are using.