How to train your chatbot … and other modern data tales

3 lessons learned from creating chatbots for enhanced customer experience

A bot has become somewhat of a mythical creature. In the early days of computing, we knew a robot when we saw one. We expected it to be a shiny chunk of metal that made cute electronic noises and spoke in stilted quaint English –like the droids R2D2 or C3PO. We fantasised about a future where robots would take over all the unpleasant tasks we didn’t like doing.

In the modern world, bots don’t always look or sound like twentieth century sci-fi movies. Not only do they predictably carry out routine and boring tasks that humans prefer to avoid, but they can also be programmed to do more complex work — faster and better than most humans. They have become so adept at imitating humans that creators give them names, companies give them employee numbers and some countries are willing to give them identity numbers too. Not surprisingly, doomsday scenarios abound of bots taking over the world.

Chatbots are an important tool to manage or enhance user or customer experience (CX). A chatbot is a software application that is meant to simulate a live conversation — this can be either a voice or a text exchange. In the CX scenario, chatbots have become invaluable in interacting with customers at various levels — from providing simple information (like ‘where is your nearest outlet?’ or ‘do you stock chocolates?’) to interacting with complex databases, and providing iterative feedback and advice. And all of this can be done with a friendly or polite tone no matter how bad a day your bot has had or how difficult your customer is. With advances in artificial intelligence and deep learning, chatbots have pretty much surpassed humans in the CX realm.

At its core, CX involves the relationship between an organisation and its customers — the goal is to create a positive experience that is built over the lifetime of that relationship. For developers and data science managers, this means that projects to create chatbots for clients provide several lessons that go beyond algorithms and other technical issues.

  1. Start with the outcome in mind: It is essential to understand the end-users and what their expectations of an online customer interface include. Customers want service providers to help them solve a problem, and knowing the industry and business processes may help to gain insights into what these problems may be. Recent surveys indicate that speed of response, friendliness and ease of use are priorities for consumers when using chatbots. Fortunately, all three of these can be included within a chatbot’s arsenal of superpowers.
  2. Stay open-minded and understand that every project is different: While you will probably do some research upfront, it is best not to start a project with pre-conceived ideas about what the solution will be. It may seem more cost-efficient to roll out cookie cutter solutions to all clients in a particular industry, but be open to adapt based on the client’s peculiar profile and customer base.
  3. The project does not end when you go live: The solution may have many iterations. Live testing generally reveals that much more work needs to be done, and once customers start using the interface, further tweaks are often required.

There are many benefits to using chatbots in CX, and consumers are more open than ever to embrace talking to a robot instead of a human — as long as they get better and more efficient service.

Make your chatbot easy to use, quick to respond, and as friendly as a few lines of code can be. And have a human or two on hand to provide extra support just in case. Because customer service is one of the hardest jobs on the planet!

Writer, explorer and lifelong learner

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store