The many advantages of a “headless” otter

It’s true, we’ve officially dipped our toes into chatbot territory with the help of Kentico Cloud. That means you can, too!

Picture of Jen Tweedie

By Jen Tweedie, Content Strategist

June 21, 2017

Share

As Content Management Systems grow and evolve (for the better!) we’ve begun experimenting with the latest in CMS tech and have learned a lot in the in the process; the creation of our virtual friend, OtterBot, is the culmination of our findings.

You’ve heard it all by now, chat assistants aka chatbots are popping up everywhere and artificial intelligence is changing the world. We’re here to tell you that there are legitimate reasons to be excited about chatbot interactions  – especially scripted (non-AI) bots – and that ecentricarts can help your business explore this option.

Introducing: OtterBot

Illustration of the OtterBot avatar, an Otter holding up the tail end of a fish
 

OtterBot is a happy and friendly chat assistant, whose goal is to guide users to useful and relevant information as specified by its overlords, ecentricarts’ content strategists. We built OtterBot with Kentico Cloud as an internal exercise to explore the merits and functionality of a headless CMS, and to get a feel for content-first web development. Throughout this chatbot journey, we uncovered the many perks of scripted bot-dom and how a Cloud-based CMS can facilitate that process.
 

Chatbots: more than just a phase, Mom

We’re huge geeks for Kentico Cloud’s user-friendly features and the bot-like capabilities we’ve been able to build with it. What’s even more exciting, is that a scripted Kentico Cloud bot can actually live up to the chatbot hype in a few key ways: 

1. Engaging and customizable user routing

The OtterBot “conversation experience” is essentially a flowchart disguised as a bot or assistant. You, the content editor, have complete control over the bot’s comments, the questions it asks, and your user’s given responses. 

This chart/user journey experience can be used to route users to a wide variety of resources in a more novel way. Based on the user’s previous responses, bots can share anything from service package recommendations (“It sounds like you could benefit from insurance plan x”), a book or article that may be of interest, a specific webpage for more information, or generate a conversational contact form – the list could go on. 

Bonus: bot questions and responses can also be randomized, which will help your content feel fresher and “smarter,” longer.

2. The Cloud is truly content editor-friendly

One of the greatest advantages of our OtterBot project is that it’s almost infinitely scalable. We’ve already repurposed our furry friend for upcoming trade shows and client demos, and completely changing, adapting, and re-routing content with the Cloud is – honestly – a lot easier than you may think.

Different content types used for OtterBot in the Kentico Cloud CMS

Kentico Cloud helps content editors understand the content entry process with good UX/UI, and no coding/development experience is required for the content editor to connect and branch OtterBot’s content and user responses. (Fun fact: I entered and mapped OtterBot’s latest conversation on the Cloud in less than two hours.)

You will, however, need a developer to set up the bot’s environment and fields. Luckily for you, we know a few capable people... 

3. No fears of corrupt AI

OtterBot can’t “learn” from user inputs – it can only say what you tell it too, so your legal department won’t need to worry about sudden insults or ideological power-trips (maybe next time, Tay)! However, with the right questions and analytics setup, you can learn about your users. 

For example, our first OtterBot iteration was geared toward large Nonprofit CMOs and shared stats and articles regarding the top nonprofit digital marketing trends of 2017. By asking pointed questions such as “Do you plan on creating video content in 2017?” and tracking responses selected throughout the conversation, we were able to gauge the “digital marketing savvy” of our users and see whether or not their journey was completed.

4. Content-first approach

Once our team landed on an initial concept, we were able to map and write the entire conversation before development had even begun. OtterBot’s functional requirements were determined by the responses and interactions that we’d written, rather than the allowances or limitations of pre-built framework. Kentico Cloud allowed us to change our project workflow by totally reversing it. 

Also worth mentioning, our OtterBot can accommodate plenty of supplementary images, .gifs and other rich media links.

 
Mobile interaction with OtterBot
Continuation of mobile OtterBot interaction
 

5. It's accessible!

In its purest form, a chat bot is just an avatar, conversation fields, and user responses/actions, so making sure each element meets accessibility standards from a development perspective is a fairly straightforward, but not yet widespread practice, as noted in this article.

When designing and developing scripted bots, consider compliant colour contrast, alt-text for the character’s avatar and any other supporting imagery (graphs, images, etc.), implementing keyboard focus states, the chat animations themselves (don’t overdo it), and how screen readers will prioritize and read the information displayed. 

For content writers, the bot’s questions should be direct, and the responses provided should, ideally, restate part of the question.

So, you think you'd like a scripted chatbot...

Great, we can help you with that! But there are a few things to consider, first.

  • How often will you be changing the chat content?
  • Who will be changing the chat content?
  • What’s the end goal of your user, and is a chat assistant the best way to accomplish this goal?
  • Lastly, “why?” is also a key question, no matter the tactic

In a nutshell, creating a scripted,“headless,” and cloud-based chatbot was a fantastic exercise in content entry, strategy, and development. We found Kentico Cloud to be fast and user-friendly, even for those with little to no CMS experience. The flexibility of a cloud-based CMS allowed an ultra-lean team to create our first chatbot in roughly 8 weeks including meetings and brainstorming, content mapping, development, revisions, and QA testing. Now that we’ve been around the chatbot block, production time has reduced drastically. 

If you’re considering a chatbot companion for your website, tell us about it! We’d love to hear your thoughts and shed some light on any questions you may have.