ePortugal

When I joined ePortugal project team, the design and comprehension phase had already taken place.
Given my business and design background, my functions in this project had extended to 3 different areas throughout the project roadmap.

The areas where I had the opportunity to work on were requirement analysis in areas such of ​​Online Certificates, UX Chatbot Journey Mapping (defining the necessary utterances and flow conversations of 3 key services) and Usability and Quality Assurance tests according to the already conceived user stories.

Therefore, the resume of the the processes regarding each of my interventions is described as follow below.

The field of online certification holds significant importance for users, often playing a crucial role during pivotal moments in their lives. Our objective was to focus on these key life events that may require certification, ensuring a prompt and efficient process.
To this end, we prioritized in our new solution the development and implementation of certificates for major life events such as:

• the birth of a child;

• the loss of a loved one;

• the registration of a marriage.

Our approach involved creating a unified master template by identifying and integrating the common elements across these certificates. This process began with the production and testing of wireframes, which evolved into both low-fidelity and high-fidelity prototypes. The system is designed to dynamically adapt based on the specific type of certificate required (e.g., birth, death, marriage), modifying the master template by adding or removing fields accordingly to suit the particular needs of each certificate.

In the current digital landscape, the increasing growth and popularity of chatbots are driven by various factors. A primary motivator is the potential to reduce customer service costs through substituting human agents with automated bots. Additionally, the effectiveness of conversational systems has highlighted chatbots as a viable interaction channel with businesses and services, complementing existing communication methods.

In light of this technological advancement, our client aimed to leverage this opportunity by implementing a virtual assistant. This assistant would significantly aid users in navigating their portal.

However, the project faced a significant challenge due to the wide range of users, each varying in their educational backgrounds and levels of speech literacy. This situation necessitated the development of a simple yet effective communication method that could cater to this diverse user base while maintaining speech coherence.

To address this, our team’s initial step was to identify the services most frequently requested by users.
We then combined this data with the intricacies of Portuguese grammar.
By training the algorithm to recognize user intentions and correlating adverbs (such as 'where', 'when', 'how', etc.) with specific terms for indexing (like 'citizen card', 'driving license', 'renewal of the citizen card'), we achieved optimal results.
This approach effectively shaped the chatbot’s decision tree and speech recognition workflow.

Quality problems that impacted the UX, beyond ease of learning, the efficiency of use or user satisfaction, are bugs and crashes that can most of the time infuriate users and impede their task completion — sometimes to the extent of demanding time-consuming workarounds.

Due to the magnitude of the project, it is normal for this kind of issues to appear.
This issue got the team running QA tests in a pre-production environment, through every browser (Chrome, Safari, Firefox, IE and Opera) in every platform (desktop, tablet and mobile) both in iOS and Windows.

These tasks allowed to prevent a great number of bugs and error that would surely undermine the system’s overall user experience. These bugs were often reported and addressed to the development team using an agile project management system.

In order to carry out the Usability Tests, we recruit users who corresponded to the 7 persona profiles created in the design thinking phase. This way, 3 scenarios were created, each with 3 underlying tasks. In each test there would be a scenario with tasks related, which the user would have to execute and express aloud the whole line of reasoning he was having at the time.

An important part of the test was to observe the reaction of users when interacting with the portal, allowing the team to see (remotely by videoconference) if they were having a positive browsing experience.