8 . 6 . 16

Rise of the Chatbots!

Artificial intelligence and technology has combined to create a new bot. It’s not Buzz Lightyear. It’s a chatbot. A chatbot is essentially a computer designed to simulate conversation with human users. They can be built, trained and deployed for a number of purposes and platforms.

Facebook sees huge potential in chatbots. Their Messenger Platform has been launched in beta and they have significant plans for it. It’s extremely likely to have been a primary factor in their recent acquisition of WhatsApp.

Emerging technology brings a number of start ups which strive to innovate and fight for supremacy. One of the most promising companies is who has just launched their service into public beta.

Using we’re going to build a basic website assistant which can be used on your website. It will be a simple chatbot which asks how you’d like to get in touch and provide two paths depending on your input. Here’s the module flow:


On chat initiation the user will be greeted with a friendly message asking how you’d like to get in touch today. In we can do this by adding a new module.


The module will have the “Multiple Choice” type with initial configuration.


It’s important to define a starting module and this can be done by ticking “Make this the bot’s start module” within the module configuration. This will ensure that the user will be served this module when initiating a chat.


Clicking “Test Bot” will confirm that this has been set up correctly.


We now want to add some logic which will push the user down a specific path based on their input. If their response contains “phone” it will display a message with the company’s phone number and closing time. If their response contains “email” it will display a message asking for their email address in preparation for the next step. To do this we’ll create two new modules – an email parsing module which will parse the user’s response for a valid email format and a bot statement which will simply display the phone number to the user.

The logic can be defined within the initial multiple choice module by defining two options – email and phone.


The connections section will define the redirection path based on multiple choice input.


When an unexpected response has been received we can configure an appropriate message for this:


We can check that the connection routes have been set up correctly by testing the email and phone paths in turn. This is for phone:


This is for email:


If you’ve entered a valid email address it will ask what message you’d like to leave.


The final confirmation message will include a thank you and estimated response time.


At this step we can add in a webhook which could send the email address and message to your business CRM.


Once you’ve tested your bot thoroughly you can get the embed code and paste it into your webpage code just after the </body> tag.

That’s it! This is a basic example but more sophisticated and useful bots can be deployed. The use of webhooks (both sending and responding) expands possibilities endlessly in regards to personalisation, data collection and any custom requirements. Zapier contains a large number of connection possibilities which will speed up development considerably when used as a webhook.

There’s serious potential with these bots. Let’s just hope they don’t become self aware.


Want to find out more about what Analytics & CRO can do for your business? Contact us here.

Matt is an Analytics Solutions Architect at twentysix with particular expertise within Google Tag Manager, CRO and all things technical such as HTML, CSS, JavaScript, Dojo, jQuery, PHP, Firebase, Angular & Ionic. He also enjoys building web / mobile apps outside of work.

Share This: