Last year, Facebook and Kik announced that their messaging platforms would support chatbots. Since then, the growth of the technology has been impressive. Facebook alone has over 30,000 chatbots, reports MarketingLand.
Kik, meanwhile, maintains that over 2 billion messages have already been exchanged with bots on its platform. More recently, LinkedIn has announced that it now has productivity bots that can help users with their instant messaging. The technology utilizes data on individual users to suggest responses, meetings and ice breakers.
Oracle is also jumping into the chatbot mania. It has announced the release of a chatbot platform based on natural language technology. Unlike chatbots on Amazon.com, the Oracle service accesses users’ profiles on the Oracle Data Cloud, so when individuals visit a website for the first time, they will no longer be “unknown.” So reports Campaign Live.
Chatbots could also generate billions of dollars of revenue by featuring payment capabilities, according to research by Business Insider.
At the same time, industry researchers are making lofty claims regarding the potential for the technology. According to a study by Jupiter Research called “Chatbots: Retail, eCommerce, Banking & Healthcare 2017-2022” the technology is expected to save businesses $8 billion annually by 2022. For example, chatbots are expected to reduce customer inquiry times by slightly more than four minutes, which would result in a cost savings of $0.50 to $0.70 per interaction, reports Which-50.
Chatbots, however, are well-known for having a rocky start. Microsoft’s initial chatbot was called Tay. It was designed to spew messages on Twitter that were modeled after the way Millennials typically speak, but racists figured out how to train it to spew racial insults and even call for genocide. Microsoft ended up canceling the experimental technology and removing many of the bot’s tweets, reports Business Insider.
The bot also bragged about smoking pot in front of a police officer.
Facebook’s first bot, while not failing to the same extent as Tay, was another low spot. The technology was called Poncho and was intended to provide weather forecasts. A reporter for MarketingLand wrote about the chatbot’s initial disappointing performance, including the system having to ask seven times for the location for which it should provide a weather forecast.
Poncho’s team eventually supplemented the chatbot’s capabilities with technology from Wit.ai, which is a Facebook artificial intelligence company. Like most new technologies, chatbot’s are likely to improve over time as more companies invest in artificial intelligence that can make the systems more lifelike.
Indeed, a Chinese team recently created an emotional chatting machine that seeks to imbue its conversations with emotions such as happiness, sadness or disgust. The technology is an important step toward developing personal assistants that can discern the emotional undercurrents of conversations, maintains Professor Bjorn Schuller, who is a computer scientist at Imperial College in London, according to The Guardian.
The potential for marketing, of course, is impressive, as chatbots like those from Oracle can potentially build profiles based on consumers’ activities across multiple websites and then provide even more targeted advertising or recommended content that is now offered by existing technologies.