In class, we’ve talked about how we are transitioning from the age of information to the age of intelligence. Companies are investing in artificial intelligence (AI) more than ever. The most obvious example would be the plethora of autonomous car companies, but even companies like Salesforce, voted most innovative company by Forbes in 2017, are implementing AI into their platform to improve their relationships with customers.
I knew I wanted to research something related to AI and I luckily stumbled upon an eye-catching headline: 2018: The Year of the Chatbot. A survey from Oracle shows that 800 of 1000 businesses wanted to implement chatbots into their services. Another article mentions that 44% of consumers prefer a chatbot to a human customer service agent. Although chatbots never live up to their hype year after year, I think they will become a prevalent part of every business as AI, machine learning, and natural language processing technology advance.
HOW CHATBOTS WORK
For clarification, chatbots are different from virtual assistants such as Amazon’s Alexa and Apple’s Siri. Although there is a lot of overlap in the technologies that they implement, their uses and purposes differ. Chatbots respond to text and sometimes voice, but virtual assistants focus on voice. In addition, chatbots are usually meant to accomplish a specific task for a business, while virtual assistants are used as a general helper.
For example, Starbucks created a chatbot that can place your order. Using your own voice or the standard chat box, you can place your favorite order at your preferred location and the bot will let you know the total cost and how long it will take to receive your order. If you are one of those people that make twenty modifications to your drink, the chatbot can still recognize that and apply it to your order. As the chatbot gathers more data and is built on a solid AI foundation, it will only continue to become smarter and emulate the process of ordering from a real barista.
Chatbots still face many challenges to emulate a human-like conversation. They receive a poor reputation from the very barebones chatbots that only provide a few services and have scripted or incorrect responses. However, the main problem lies in the fact that it is very difficult to create a machine with the complexity and intelligence of a human.
TECHNOLOGY USED IN CHATBOTS
- Artificial Intelligence: This field of study focuses on the theory and development of machines that are able to perform tasks that require human intelligence, such as decision-making and speech recognition.
- Natural Language Processing: This is an area of artificial intelligence that deals with the interactions between human languages and computer languages. The programs are meant to gather a large amount of data to analyze, understand, and derive meaning from it.
- Machine Learning: This area of computer science deals with designing programs and algorithms that teach a machine to perform a task and improve upon the task as it is used more. With all of the data gathered from natural language processing, machine learning will allow chatbots to find patterns in in the data and make predictions based on the data (AKA provide meaningful answers to user input). The development of solid machine learning technology is what will set mediocre chatbots apart from the chatbots of the future.
FACEBOOK CHATBOT PLATFORM
The launch of a platform for building chatbots on Facebook Messenger was a key event in reviving the craze around chatbots. In the first year of the platform, developers had built more than 100,000 bots. David Marcus, Vice President of Messaging Products at Facebook, said that the messages between businesses and people had doubled to billion messages per month. This amount of data collection/business-to-customer interactions can turn these chatbots into powerful marketing tools.
One advantage of chatbots is that they can be integrated into many messaging apps, WeChat and Facebook Messenger being two of the biggest. Besides offering basic interactions, chatbots can also conduct marketing surveys or distribute new content through push notifications.
Chatbots can be used for internal business operations as well. An article on Forbes also suggested the idea that chatbots can be used to manage marketing expenses. After supplying the chatbot with enough information on a company’s budget, it can perform tasks such as alerting a team member if they exceeded their budget limit. It can also analyze data and make suggestions based on processing the massive amounts of data. Again, this is all dependent on building solid technology, which is something that can be achieved.
EXAMPLES OF COOL CHATBOTS
Although this is a platform for building AI chatbots within the Salesforce suite, this definitely deserves a mention. Einstein allows customers to build a personalized chatbot that offers business insights through data analytics, all through the reliable Salesforce cloud. In addition, customers can build a typical chatbot that engages with community members to provide the right content at the right time.
Developed by Stanford psychologists and AI experts, WoeBot is a 24/7 mental health chatbot that specializes in cognitive behavioral therapy. WoeBot tracks a user’s mood and finds patterns in the responses to share information about the user. The chatbot is programmed to inform the user that it is not a substitute for professional help, but it can potentially be a helpful form of therapy, all through the use of Facebook Messenger. Although it acknowledges that it is not a human, WoeBot provides very human-like responses.
Ridesharing became even easier once Lyft introduced a chatbot that allows you to request a ride through chat or voice. The chatbot provides you with driver details and ride locations. In Slack, you can even use a slash command to request a ride, letting you commute from work to home without even pulling out a cell phone.
In my opinion, developers have not even come close to realizing the full potential of chatbots. As research into AI grows and machines become smarter, chatbots will be able to fulfill a variety of tasks in a human-like fashion. 2018 may very well be another year that chatbots have not lived up to their hype, but there is no doubt that current chatbots are already saving companies money and time through automated processes and data analytics and increasing customer satisfaction.