PORTFOLIO CHATBOT PROTOTYPE.

phone new.png

Backstory. 

During my training as Voice Interaction and Conversation Designer at the Digital Assistant Academy, I was tasked with coming up with my own chatbot prototype for my first Capstone Project.

 

I wanted a way to develop my skills in chatbot design whilst also showcasing my own personal portfolio and career background within a conversational model.

It can be difficult to find interesting information through people's portfolio.

It can be difficult to show a personality in a resume.

It can be difficult to chat to people throughout busy lives.

So why not create a chatbot that can solve these issues in one simple conversation platform?

Platform:

Originally Facebook Messenger, however it would ideally be integrated onto portfolio website.

Tools: Conceptboard, Botsociety.

Conversation Flow Mapping & Sample Dialog.

  • User data and bot usability test score.

  • Usability report.

  • Chatbot adjustments.

Key Processes.

  • User research to identify and prioritise features.

  • Conversation flow maps to visualise the flow.​

  • Prototyping for Facebook Messenger chatbot

  • Usability testing.​​

Research.

Initial research suggested that the personality and language style should be formal with occasional mix of the friendly aspect added in to give it a human element, whilst also staying professional. The majority of people liked the chatbot idea with 20% either unsure or wouldn't use it, and that it would be better suited for creative industries over more formal industries. However it could be popular within the recruitment process, to save time asking simple questions about a particular candidate.

Top 5 features users wanted to be able to find out:

  1. Identify main skillsets.

  2. Most recent career jobs.

  3. Recent projects.

  4. Short bio or background information.

  5. What technology/software he uses.

Problems encountered.

  • Bot was fairly ‘linear’ in is design approach. -  moving forward, I would allow for user to type in what they want, creating a more conversational approach.

  • Due to it being an information bot, choosing how much information to include without it being too much for the user proved difficult

Comments from testing.

  • The messages were too fast, couldn’t read them quick enough.

  • Unable to scroll up or down to view previous messages.

  • A few minor path errors and inactive buttons.

  • Messages were too long, must be more concise and direct.

  • Fun, engaging idea.

  • Users preferred clicking buttons rather than typing.

  • Too many options to view, bit overwhelming/confusing to choose.

  • Reduce unnecessary words that don’t add much to the conversation.

  • Liked the hyperlinks to the portfolio.

Screenshot of selection in Botsociety prototyping tool.

Improvements.

The improvements were made in relation to the first stage of user testing. By taking a look at all the comments I was able to make the following adjustments:

  • All errors were resolved throughout the design to ensure a complete flow around the chatbot.

  • Parts of the wording were changed to be more concise with the information displayed.

  • More buttons were added throughout the design to create an easier experience for the user.

  • Additional confirmation quick-links to give the user more control of the chatbot speed.

  • Ensured all paths are simple and easy to follow.

Moving forward.

 

The participants were reasonably satisfied with the Portfolio Chatbot Prototype, finding it fun and interesting to use. There were a few concerns about the length of time it took to move around the chatbot, but with the alterations and revisions indicated on this page, it is hoped that this will significantly reduce the time spent to find information. Further testing would be needed after these alterations, to ensure that improvements are made.

The inclusion of natural language processing would be included to establish a more natural conversation and less of a linear conversation path. This would then be constantly monitored to improve the machine learning capabilities.

However, to move move forward with this chatbot, I must first re-evaluate its original use case. Does it solve a problem of speed and efficiency? Probably not in its current state. But does it provide a fun and engaging idea? Yes it does. That being said,  it has been a great project to be involved in. Making errors and improvements only adds to the learning experience, whilst I understand the main processes of conversation and chatbot design.

Have a try yourself by clicking below to try out the prototype yourself. You may even find out a thing or two about me!