Build a Chatbot using Snatch Bot

Now that you have a portfolio and your website, you may start thinking about adding a chatbot to collect email addresses and answer basic questions. Follow along as I try the Snatchbot.

The Dashboard

After logging into the platform, you are taken to a dashboard. Where your current chatbots are kept as well as data on messages exchanged and messages by channel, to create a new bot, select the create a bot in the upper left-hand corner of the screen.

Your First Chatbot

After selecting a new bot, you are presented with several options for templated bots. Select a blank bot if you want to design everything yourself or use a template if you’re going to modify a pre-existing chatbot. You can add an image as well as give your bot a name and a description. Then select create.

Adding Interactions

In the next image, notice that there are two sections; add a new interaction or plugin and chatbot scheme. The chatbot scheme is a preview of your interaction after you build them. It is also an area where you can test your interactions. To add an interaction, select add new interaction just below your bot’s name. There are several YouTube videos uploaded by SnatchBot, as well as a robust set of instructions.

After you select to add a new interaction, you will have multiple options. A chatbot statement is the most common. Here you can make a statement or ask a question and anticipate some response. You will then add connections that are commonly if-then statements to lead to the next action. You can also choose to extract emails from the user response, extract a URL, a phone number, a date, a number, address, a duration, a JSON API, a search interaction, await a response from the user for a human or a translation. To build this simple chatbot, we will only be using the chatbot statement and email extraction interactions.

After selecting a chatbot statement, name your interaction to make it easier to keep track of your network diagram. You will add the chatbot’s message. This message is the statement the bot will make when the user reaches this part of the communication. Since this is the first statement in the network, I would like to introduce the chatbot and ask the user if they would like to chat.

Add A Connection

After you add the chatbot’s message, you will need to connect the message to another action. I have added two additional interactions to my bot to prepare for this step. In my first chatbot message, I asked, “Would you like to chat?” There are two possible responses, yes and no. If the user says no, then I would like the bot to say, “Thank you for visiting my website. Please let me know if I can help you in any way.” If the user says yes, then I would like the bot to say, “May I have your email address in case we get disconnected?” So the next steps are to set up two simple connections depending on whether the user inputs yes or no—select “connections” at the top of the screen. Then select “simple connection” near the bottom of the page.

If the user’s response to the interaction is no or any form of no, we would like the bot to respond with the goodbye response. To do this in the first box under the simple connection, select “response to this interaction.”

The user response could either “contain or not contain” a word depending on how you want to set up the connection, so the greatest number of user inputs are included. If you select “contain,” it will look for that word or any synonyms for that word you add. If you choose “does not contain,” it will move to that action as long as the selected word is not contained in the message. Using the “does not contain” option covers many more possibilities.

Next, you will enter the keywords that the user response either does or does not contain. Make sure you add the keyword in both capitalized and lower case forms. Type the word and then enter. After you enter a keyword, you have the option of searching and then selecting synonyms for each keyword.

The last step in building the connection is choosing which interaction the response will lead to.

Test You Chatbot

After adding several interactions, you may want to go to the bot scheme area to test your bot and look at a diagram of the connections. Select the conversation bubble in the greeting interaction to bring up the test chatbot window.

Sample Chatbot

You can make your interaction as complicated or as simple as you desire. In this case, I used simple if-then variables, ultimately trying to get the user’s email address so I could contact them to determine their needs and answer any questions. If I were selling a product, I could answer questions on the price, payment, and return policy just using the chatbot. Those with more advanced coding skills can use artificial intelligence and coding to build a chatbot that mimics conversing with a person.

Deploy The Chatbot

The next step is to deploy the chatbot to the channel from which you would like to collect data. Select “Channels” in the left-hand column and then “about.” In the about section, you can provide a name for the bot, a custom signature, your contact information, including your website and your email.

In the next tab, “customize,” you can choose the styles and color scheme for your chatbot so that it matches your website design. Finally, under the “get embed code’ tab, you will get the necessary code to embed the bot on your website. You will need to follow the instructions on how to embed the code found on the Snatchbot website.

The easiest way to deploy it is to add the Snatchbot plugin. Navigate to your Word Press website, and select add a new plugin. Search for Snatch Bot. Activate the plugin.

Snatchbot Plugin

After activating the Snatchbot plugin. Sign in to your Snatchbot account and select your bot to automatically deploy it to your website.