Everyone is talking about chatbots, but they are still at a very early stage of their evolution. For the vision to become reality, a measure of added value is necessary.
When people mention the word “bot” today, most of them are thinking about social bots. These publish predetermined messages on social networks, create new ones by themselves, respond to keywords or automatically share content on related topics. They were heavily used during the recent US presidential election campaign. As exciting as this topic is, however, today I am focusing on another type of bot: the chatbot.
A growing range of possibilities being met with skepticism
Shopping bots, media bots, service bots, even personal bots – these digital agents are now springing up like mushrooms. Some of them are more curious and others less so. Take for example the “Oshi-el” system by the Japanese company NTT Resonant, which offers the user advice in love and relationship matters – an agony aunt for the broken-hearted. Coffee junkies, meanwhile, can look forward to a digital barista; Starbucks is testing a new service within its app that will enable users to pre-order coffee by voice command or chat.
It is in the service sector that chatbot-driven services can prove a real asset. Last month, my colleague Michael Strauss posed the question in this context of whether chatbots would replace apps and websites. For now, however, the answer must remain tentative: it will not happen today or even tomorrow. Many users remain skeptical, the three main arguments being:
- People don’t want to communicate with a computer.
- Users doubt the reliability of chatbots.
- The technology is not yet mature enough.
Chatbots are not a new invention
Looking back, it becomes evident that chatbots have a surprisingly long history. The first one was Eliza, a virtual psychotherapist, back in 1966. Even Microsoft’s Clippy is just a chatbot – if also a rather simple predecessor to the current concept. Today chatbots are more hyped-up than ever, and this is due to the huge availability of messengers (such as Facebook Messenger, Skype, WhatsApp and so on). The providers of these services have recognized the potential for chatbots and are increasingly creating APIs for the development of these digital agents. A salient example here is Facebook. From its start in May 2016, users can now access over 30,000 chatbots within Facebook Messenger. An example of this is an offering by the Kayak travel portal that enables users to plan their journey entirely via Facebook Messenger.
Convenience – but with pitfalls
The future of messaging: the user chats with a virtual assistant that helps them with their everyday activities. An example of this is Amazon’s Alexa. The online retailer also offers the Echo Dot, a speech-controlled device that connects to the Alexa voice service to play music and to deliver news, sports results and weather forecasts. It also allows online purchases to be made by speech command. Sounds like child’s play – and it is, as a family in Texas was soon to discover. They were taken by surprise when they received an order confirmation e-mail for a dollhouse and a packet of Danish butter cookies. The solution to this riddle: their six-year-old daughter had been talking about cookies and dollhouses near the loudspeaker, and unbeknownst to them, this had placed an order.
Added value makes the difference
The examples given above are an indication of the direction in which the chatbot scene is likely to develop in the future. For now, however, its potential remains palpable rather than visible. I see three main sticking points here: voice control, artificial intelligence, and the systematic expansion of meta-chatbots.
1 Voice control
While the World Wide Web is already awash with text-based digital assistants, there is a significant need for speech-based versions to catch up. To make communication between human and chatbot even easier and more convenient, in my view, voice control is a basic prerequisite.
2 Artificial intelligence
Achieving a natural style of conversation still seems a long way off. The requirement for this is a self-improving bot equipped with artificial intelligence that learns from its history and from external data sources, and that solves problems autonomously. Only at that stage will there be few barriers remaining in the way of digital assistants.
Currently, it is tedious for the user to have to install a separate app for each application and each product and to create a new user account for each one. For this reason, I see few prospects for isolated chatbots within individual apps. What is needed are meta-chatbots that can combine all functions and services on a single platform – whether that be Amazon, Apple, Facebook, or Google. Only then will chatbots be able to function as personal assistants in every area of the user’s life.
Only when these three elements of added value are realized will chatbots become established in the mass market. And that, for sure, won’t happen today or tomorrow – but will surely do so the day after tomorrow.
Picture credit: Shutterstock