Continuing our ecommerce trends 2017 series, here's the low-down on chatbots.

What are chatbots?

Chatbots are becoming an increasingly common sight throughout the ecommerce world. Specifically designed to replicate human interaction, chatbots exhibit intelligent behaviour like that of a human to convincingly simulate how a human would respond as a conversational partner in the given circumstance. A user should therefore be able to message a brand or business in the same way they would message a friend.Chatbots are often powered by both rules and artificial intelligence (AI) involving natural language processing systems or machine learning – meaning they may understand language and not just commands, and can get smarter as they learn from conversations.Rather than just a question and answer format (found in virtual agents – a concept that has been around for a few years that simply matches keywords with answers) a chatbot relies on increased intelligence allowing it to change the subject, suggest related topics, and even demonstrate emotion based on the responses given by the user.

Why are chatbots gaining in popularity?

Chatbots are often used to deal with the ever problematic issue of providing useful customer service answers, without the time spent and cost incurred when dealing with (human) customer service agents. That said, the AI technology being utilised in chatbots means they can be used for a variety of other purposes.One of the main reasons for the increase in brand usage of chatbots could well be Facebook's involvement. During last year's F8 summit, Mark Zuckerberg announced the integration of chatbot technology into Facebook Messenger. Since then brands like Uber, Skyscanner and Amazon have already got in on the Messenger chatbot act. Why use Facebook Messenger? It's a ready-made chat platform that over 900 million people already use, and therefore don't have to download specifically for the task. It makes more sense to use it than to not.

Who's using chatbots?


HealthTap is an interactive health company that lets users ask questions, share photos, get test results and referrals from doctors. HealthTap's chatbot utility uses Facebook Messenger. First, the chatbot analyses requests and then shows similar questions asked by other users. Users can then see how many doctors have offered answers to each question and check out each response. This particular chatbot example combines human and computer interaction. If a suitable answer is not established to a user's question, it offers to send the query to its network of real doctors.


The Uber chatbot uses the Messenger app and lets rides be requested by starting a conversation with the chatbot. The chatbot can then provide the user with a fare estimate, the driver's name and registration, and what time they'll arrive, all within the same Messenger app, removing the need to load a separate app altogether. Handy features such as clicking on an address that's been sent to a user by a friend to request an Uber, adds further to the automation.


Duolingo have developed AI-powered messaging chatbots for language conversation practice. The chatbots allow language learners to practice their conversational ability in their chosen language with a chatbot rather than a human, which is both less embarrassing (if mistakes are made) and much more convenient than say, hiring a tutor or going to a foreign country. Users get to practice language conversation on-demand, whenever convenient. The Duolingo conversation chatbots respond to thousands of different answers, and conversations are structured around different scenarios and topics. Some chatbots even have different personalities.


KLM gets on average five questions a minute via Messenger (13 messages a minute during peak times) and so a degree of automation is no doubt of considerable advantage. KLM adopted a slightly more human approach, opting for a chatbot-assisted agent – seen as the next step up from a virtual agent. It works with the same type of intelligence or 'smart data' (suggested answers are based on historical data collected on each customer) but with human agents working as quality control to ensure a chatbot’s suggested answers are correct.Karlijn Vogel-Meijer, KLM Social Media Manager (speaking at Festival of Marketing 2016), said that the approach gives "the best of both worlds - a timely answer, a correct answer, and a personal answer. The best of humans and the best of tech".

Sources and further reading... - How KLM uses bots and AI in 'human' social customer - Practice Your Spanish (French Or German) With Duolingo's New - Five pioneering examples of how brands are using - An introduction to AI and customer - The Complete Beginner’s Guide To Chatbots