Listly by crawfordcopy
As Chatbots increase in popularity, there uses have become more diversified. One area in which Chatbots have long been popular is e-commerce. When you look at it, it’s not surprising.
The online marketplace is seriously competitive, and highly intertwined with social media. Chatbots are able to bridge the gap between the two, and bring a powerful e-commerce experience to the channels where potential shoppers spend this time.
Once your build is complete and tested, it’s time to launch through the channels you selected earlier. This may seem like the end, but it’s really just the beginning.
The Chatbot you launch on day 1 should be very different on day 7. During that time, you can see how users are or are not interacting with your build. Based on this, you can make adjustments on the fly to improve areas where failures are occurring, and optimize flow through the conversation.
All in all, building a shopping Chatbot is only as difficult as you make it. If you put the effort into outlining what you want and choosing the right platform, you can be up and selling in no time.
How you build the Chatbot depends on whether you’re choosing to code it yourself or use a build platform like SnatchBot. For a step-by-step breakdown of how to build a Chatbot on SnatchBot, you can refer to my guide here. In this exercise, I created a simple yet useful Chatbot you can have up and running in minutes.
As you proceed through your build, there are a few important things you should be doing.
First, test as you go. There are a few reasons for this. Testing your Chatbot as you go helps you understand the flow of the conversation. If you wait until the end to notice that things are off at the beginning, it can create a lot of extra work. Similarly, testing along the way helps you isolate any errors or missing connections. Rather than getting to the end of your build and discovering that there’s an error somewhere in the many modules you created, testing allows you to isolate them with ease.
Second, if possible, get input from others, especially potential users. It’s easy to start building a Chatbot and think only about how you would use it. Sure, you’ll always be relying mainly on your own experiences when you create something, but it’s helpful to hear the opinions of others. This is especially true for Chatbots, where one small change at the start of the conversation flow can have a major impact on how the rest of the build works.
There are a number of ways that you can go about building a Chatbot. Choose yours based on your technical abilities, as well as what you’re looking for your Chatbot to do.
The first way is by hand with code. This is the most difficult and time-consuming way, as it usually requires a strong knowledge of a few different coding languages. If this is what you’d like to do, platforms like Facebook and Twitter have provided sample code you can use to get started. The upside to this approach is that you’re able to have more control over what you’re building. However, given that what you’re building will ultimately display in a pre-set social platform, what you can customize is relatively limited.
The second and most likely better option is to use a Chatbot build platform. These software platforms make it simple for non-coders to create and deploy a great Chatbot.
There are a number of these platforms out there, but for my own builds, I consistently choose SnatchBot. I make this choice for a few reasons:
1) It’s simple and intuitive, meaning you can get up and running in minutes.
2) It has a lot of powerful features, ranging from built-in NLP to customizable delayed responses.
3) It has a stellar suite of analytics and measurement tools.
4) You can easily build a Chatbot that can switch between channels.
Essentially, it has whatever you need for whichever kind of Chatbot you’re building. In my case, it makes it easy to quickly and easily set up a Chatbot I can link to my businesses Facebook account. From there, I can begin reaching my goal of selling to my customers through this social platform.
Next, start to think about where you want your Chatbot to come to life. Generally, there are 3 main areas where Chatbots exist:
1) Online: Web-based Chatbots primarily function as customer service assistants. You’ve likely encountered them on Chat functions of websites.
2) Social media: These Chatbots are launched on public-facing accounts of businesses, organizations and more. They can exist either within the messaging channel of the platform, or as a part of the main content delivery system.
3) Internal platforms: Chatbots are now being deployed within internal tools like slack. These Chatbots are mainly focused on improving productivity and simplifying tasks and communication.
In the case of my coffee shop, I’ll be looking to social media. This is primarily where I reach customers, and would like to begin selling to them there as well.
Now that I have the overall channel selected, I can move on to examine individual platforms. From Twitter to Facebook Messenger to many more, there are a great variety of social platforms to choose from. When making this selection, consider:
1) Which channels the audience you’re hoping to reach uses most frequently.
2) Which channels have functionalities that best suit what you’re looking to do.
On the first point, you can figure this out by looking at your existing audience and seeing where they frequently post.
On the second point, take a little time to explore your options as different platforms often have different specialities and functionalities. For example, Facebook Messenger has very powerful e-commerce abilities. Twitter, on the other hand, is geared more towards customer service.
For my coffee business, I’d select Facebook Messenger. First, Facebook is primarily where I interact with my audience. I have a strong following, and would like to start selling to my customers in the same place where I’m speaking with them. Second, as noted above, Messenger is an e-commerce juggernaut in terms of Chatbot functionalities. It can not only process payments easily, but it’s also well built for displaying high-quality images in an easily navigable format.
Now that we have the scope of our Chatbot and where it will exist determined, it’s time to look at how where we’ll bring it to life.
The capabilities of Chatbots cover a very wide range. At their most basic, they are a few lines of pre-written text. At the other end, they are AI-powered pieces of software capable of instantaneously conversing with a human. It’s important that before building, you learn where on the spectrum you want your Chatbot to exist before you start building.
To determine this, ask yourself a few questions:
1) How much time and money is available to put towards the project?
2) What is the main goal of the Chatbot?
3) Will it be driven by buttons, conversations or both?
4) Will it require AI and NLP?
5) Will a human be able to respond if necessary?
By asking yourself these questions, you’ll be able to start putting together a more solid idea of what the finished product would look like. If these questions are difficult to answer, check out other Chatbots and see if there are any that are similar to what you are looking for. You can test them and see what kind of functionalities are available. A good resource for finding these Chatbots is the SnatchBot Store.
To relate back to these questions, let’s pretend I want to create a Chatbot for my online coffee store.
My time is somewhat limited, and it’s not something with a major budget, so I’ll likely stick to free services.
I want it to provide e-commerce options so customers can shop through social channels.
Given that it will be product-driven, it will rely mostly on buttons and images.
No major AI will be needed, but a few trained keywords would be useful.
Yes, I will be available if a user encounters a problem that requires assistance.
From here, I have a solid understanding of what I’m building. This prevents the scope from slipping away, and allows me to move forward with a solid plan. Starting to build without answering these questions means that you can miss out on key functionalities, or build without really knowing where you’re going.