What I learned configuring a chatbot for fun. | Dow Goodfolk, brand strategy, identity, packaging and digital design agency
Dow Goodfolk
Back to our news

What I learned configuring a chatbot for fun.

What I learned configuring a chatbot for fun. image

Let me start by saying, I don’t code.

May 13 2019


I’m an account manager, my job is to help provide solutions for our clients and to work with internal teams to help bring those solutions to life.

With that being the case, I have to learn and understand digital products as much as possible without constructing them myself. Over the years I’ve developed my knowledge by being present and involved in the process as much as possible. Asking developers countless (no doubt annoying) questions, or diving in where I can, like helping create or manipulate excel databases (tedious at times, but now I understand a lot about data structure that I didn’t know before) and learning on my own. There’s so many things you’ll discover just by being curious.

It’s this sort of learning that I feel is invaluable to understanding digital and all the subtle nuances that the ‘doing’ teaches you. I’ve been curious about chatbots for a while but my knowledge was limited to a general understanding of them. I wanted to know if I could make one myself. Not being able to code, turns out to not be a problem. At least not at my level of need. 

Enter Landbot.io. Landbot is an incredibly easy to use tool that allows you to create your own chatbot from scratch without needing a developer.

The basic function of a chatbot is simple. It’s a program that receives a message from a user, reads the message and then provides a meaningful answer. Think online shopping and asking a chatbot if they have your favourite chuck taylors in a size 6. A simple answer such as, yes and here is the product page, would constitute a simple but meaningful engagement.

When making it yourself (without using a true AI that can learn for itself) the challenge lies in the shear volume of possible interactions and the nuances of basic conversation. You have to understand the plethora of potential user needs and then write a script or journey for all of those if you want to create a meaningful chatbot.  

 

For example;

Someone asks if your website has pink chuck taylor originals in size 6

The program checks the website and finds out either;

No, there are no pink chucks in size 6, but there are pink chucks in sizes 6.5.

Or No, there a there are no pink chucks in size 6, but there are blue, green, black and white ones.

 

The logical answer is no, but a meaningful answer should provide a more meaningful outcome. Like “sorry we are out of stock, but have options in blue, green, black and white. Would you like to take a look at these?”

So, in order to create a truly engaging chatbot experience you need to have a deep understanding of your product or service and the ability to anticipate and interpret all (or as many as possible) of your customers pain points and needs. You need to understand how they think and how to construct meaningful answers.

I guess the most interesting thing, although unsurprising, that you need to know when creating a chatbot is your customer. And what they need from you.

Design at its best is customer-centric.