EKO NEURAL

Web solution

“We want all kinds of people to be able to create a chatbot easily and without specialized knowledge in artificial intelligence; this from his area of ​​specialty: doctor, architect, designer, psychologist."

  • My contributions:
    Product discovery
    Analysis
    Research
    Navigation flows
    Wireframes
    Mockups
    Prototype
    User tests
    Interface design
    Layout
  • Tecnologies:
    UXPin
    Illustrator
    Visio
    Visual Studio
    MVC project
  • Client:SICOSS group
  • Metodology: Scrum
  • Year: 2018

Context

Currently all kinds of companies have operators that are dedicated to the personalized attention of specific topics for users. These issues are usually repetitive and the particularities are addressed by people of higher ranks.

Unfortunately this type of attention is limited by a schedule and even during the 8 hours of workday, there is not always the immediate and precise attention that is expected.

We must take into account that users are now much more demanding than just a couple of years ago, so an immediate, personalized and quality attention will take the lead over the competition.

Objectives

  • Attention 24/7

    Stay available to the client even on weekends.

  • Clear and immediate information

    Offer clear, immediate and always accurate answers according to the context of each user.

  • Custom reports

    Take into account data that the user can provide to perform calculations with accurate results.

Proposed Solution

With the help of IBM Watson, generate a web solution that allows all kinds of people to create a chatbot and train it as much as they want, to assist their users on general and specialized topics.

Users

Administrators of the tool and all kinds of people and organizations that need to provide information about a topic of their knowledge.

Usuarios

Meet Erik

36 years - Businessman (Nail Spa)
Use of digital tools: intermediate level

Erik is a very creative, dedicated and empathetic entrepreneur with his clients, always likes to listen to what they need or would like their business to have to grow and be successful.

I would take advantage of this tool so that its users can request at any time reports about the services and schedule appointments.

Usuarios

Meet Laura

41 years - Human resources director
Use of digital tools: basic level

Laura is a perfectionist and dynamic person, she always hopes to give the best of herself in her work and that the service of her department is impeccable.

Laura can take advantage of this tool to make her staff more efficient and the chats can resolve employees questions and new income people.

Usuarios

Meet Mariana

31 years - Event organizer
Use of digital tools: intermediate level

Mariana is a very creative person, she always likes to renew herself and improve her services.

A chatbot would be very useful to forget about giving reports and only dedicate to what she likes, organize the best events for her clients.

System flow

From several sessions of understanding the whole process with the client, the project manager and development team, I elaborated a flow that shows in detail all the components of the solution:

  • Customer Administration
  • Project management
  • Chatbots training
    • Questions
    • Intentions
    • Entities
    • Affirmations
    • Formulas
  • Chatbot tests

This material was key to start with a rough estimation as well as to have clarity of the components to be considered within the MVP phase.

EKO-NEURAL

Wireframes

Part of the flow begins with the rough sketches of each of the components until reaching an initial prototype.

The results yielded very valuable information:

  • It`s required a previous workshop clearly structured and with basic exercises related to the turn of each company to start a chatbot training that they can continue working with greater security later.
  • In several of the functions of the system, informative tooltips are required to confirm to the user what each action refers to outside the universal context.
  • It`s necessary to expose the client the amount of questions with their corresponding statements that has a chatbot in each level of expertise.

Mockups

Finally I managed to reach the stage of visual magic, adding styles of colors, typographies, dimensions and images; all this based on the identity and logo already defined by the business area.

In the second stage of tests with users, interesting results were obtained:

  • They identified many of the elements much more easily thanks to the assigned chromatic priority.
  • They were very easy to do all the training of a chatbot from a single screen.
  • However, some screens in the training section still had changes.

Takeaways

Análisis
Analysis

This solution presents two main ways to train a chatbot, both involve considering at least 4 of the 5 components for a question. My mission was to analyze and generate that logical order with different alternatives to be able to create in a fluid way a question with all its components.

Empatía
Empathy

Being an entirely new tool, empathy played a very important role for me. Analyze all the possibilities and paths that users could take was an interesting challenge that consisted in several iterations of tests with the users, until arriving at the ideal solution.