Chatbots, the UI of AI
Conversational agents are the new human computer interface for interacting with software and the power of the internet, in many cases replacing mobile apps and the internet browser.
"In 2007, I had this vision when I first touched my very first iPhone where I kind of understood what the next five years would bring, and I haven't had that kind of clarity since... and now, I have the same kind of feeling about bots, about conversational UIs." - Phil Ebin, Founder of Evernote
the chatbot army
In many parts of the world, people are using text more to communicate quickly than ever before. Whilst private text based instant messaging chat rooms were common in the early days of the internet, and similar concepts have had occasional successes since, only since Facebook, Telegram and WhatsApp provided private multi party chatroom have they become a mainstream feature of sharing holiday photos and co-ordinating events with family and friends.
The mobile messaging platforms have dominated the text messaging market, overtaking SMS in 2014. the text messaging market, overtaking SMS in 2014. The "instant message" style of interaction is now starting to gain a foot hold in how we interact not just with each other, but with computers. It's no surprise that the same popular consumer messaging platforms all now support “chatbots”, and that bots are increasingly added on to proprietary websites and apps.
Voice recognition has also been making fantastic progress in the consumer psyche over the past 18-24 months, and has been implemented by all the"Big 4" consumer tech platform providers as Amazon Alexa, Google Home, Apple's Siri, and Microsoft's Cortana. People are becoming accustomed to using voice to instruct AI personalities, and to listen to an audio response.
The natural evolution of the medium will be video, where the same technologies can be used to emulate a video conversation with a bot (or its Avatar).
Ultimately chatbots help businesses and organisations improve the experience of their customers and employees by improving customer communication channels, be it through video, audio or text.
"By 2020, customers will manage 85% of their relationship with the enterprise without interacting with a human." - Gartner 2017
winning over the consumer
Increasingly, enterprises are offering a “chatbot” mechanism for consumers to interact with them. This isn’t the same as “live support” - although its sometimes hard to tell! A single AI can simultaneously deal with an infinite number of customers simultaneously, will behave more consistently than humans, and can service customers faster. In many situations they will also be able to authenticate existing customers instantly, via the decentralised authentication mechanism of the chat platform, e.g. Facebook.
Chatbots are a form of AI tech, focused on a narrow problem domain, and enabled by advances in natural language processing technology combined with machine learning. Natural language processing, or at least the perception of a user that they are interacting with a real person, is also a key element of what makes up an AI solution, it is a critical portal for an AI solution, which is one reason the market is estimated to expected to reach USD 1.25 billion by 2025.
Its been shown that consumers like this and want more of it. Some statistics from Aspect Software state 61% of consumers think that chatbots allow for faster resolution and answers. 65% also feel good about themselves when they avoid talking to a live person to resolve an issue. One supposes, there is no guilt in leaving half way through a dialogue, for a few minutes to make a cup of tea, and then continuing as if nothing had happened.
The use of a chatbot in a customer sales or support context is fairly transactional, the customer reaches out to the AI with a specific need or intent, usually to get information, but perhaps also to provide or update information.
"Five years ago, we reached a tipping point in digital adoption, where the majority of people (51%) would prefer to transact online than deal with a person. Today, we are seeing similar attitudes emerge towards AI; almost half (49%) feel comfortable dealing with chatbots and a third actively prefer them for basic pricing and availability queries." - Colin Smith, Managing Director, Customer Science
In addition to accelerating customer interactions, AI is also far cheaper per transaction, and human customer service staff can focus on more complex interactions and issues. AI also is constantly available, unless it is compromised by legacy systems it needs to interface with.
"By 2022, one in five workers engaged in mostly nonroutine tasks will rely on AI to do a job." - Gartner 2017
transforming the enterprise
An even more complex and interesting concept is the deployment of chatbots in the workplace. Enabling a worker with AI technology can provide numerous valuable transformational benefits.
- Information retrieval from common enterprise sources such as calendars or email will become more common in 2018, asking for the location of your next meeting or last week's presentation to be emailed to you will become normal. You can already link easily link Alexa for Business to some cloud services, e.g. Office 365.
- The AI can act as a knowledge store and help fulfil standardisation and digital transformation goals within an organisation. A key part of implementing AI is defining the organiational body of knowledge, and structuring this around the goals and objectives of employees.
- Modern knowledge workers are constantly distracted by emails, phone calls and other digital interruptions. A key feature of using an AI should be the ability to query and ask questions which require data to be derived from multiple sources.
To achieve change that will be a real shift in thinking, the AI behind the chatbot "user interface", needs to be fully aware of the workers goals, it needs to understand them empirically, and be able to interface with the varied work-flow and information systems in the enterprise.
Once an employee is managing their workflow through an AI, you can more easily replace manual activities with automated ones. You don't need to retrain people on an obscure process if it is taken off their hands and handled by robotic process automation, presenting as a chatbot conversation to the employee.