top of page

Atlas | Independent Honours Thesis | 4 months

Digital Secretary: Enhancing productivity of home-based small businesses through user experience design

Process

  • Discovery

  • Interviews and real-users testing

  • Hi-fidelity prototype

Results

  • Integrated the use of AI chatbot as the key focus

  • Improved desktop interface that is well-received from users

  • Discovered that users distrust AI chatbot and will need time and assurance to fully rely on AI services

Back in 2020 when the pandemic took us by a storm, business models were forced to change with new policy enforcements, and rapid digital transformation ultimately changed the way we work. With the rise of hybrid work arrangement, many Singaporeans had greater flexibility in turning their side hustle into a home-based small business (HBB).


Although there are many apps available to help with increasing productivity and organisation, it is uncommon to find an app that will allow home-based business owners to combine both their operations and marketing procedures in one.

Competitors' Analysis

I conducted a competitors’ analysis on four task and business managing related apps — Cococart, Evernote, Microsoft To-Do, and Todoist, as well as two social media business apps — Meta Business Suite and Instagram for Business.

Key findings:

1. Users want an app that has effective task organisation functions that they can customise to best suit their needs.
2. Users want to enjoy greater convenience and increased productivity through unique features.
3. Users want the app to have a low learning curve and be simple but effective.

User Testing
2 males, 3 females, age range 19 - 31
Key focus areas

1.    Ease of navigating through the site.
2.    Clarity of functionalities of the site. 
3.    Usefulness of functions available to target users.

Results
Naming of 'Today' menu in sidebar can be improved.
Dashboard Landing - Monthly.png

Updated sidebar menus

'Task Manager' page was not as useful as expected.

Decided to streamline the L1 menus further by removing the 'Task Manager' page and incorporating the task management features into the Overview dashboard.

A simpler way to manage discounts is needed.

Separated the discount tab from the add product flow.

My Shop.png
Discount 1.png

Updated discount features on 'My Shop'

Maintaining the statuses of each order is too tedious.

Streamlined the statuses available to only 'Pending', 'Approved', or 'Rejected'.

My Orders Landing - order date.png

Updated status features on 'My Orders'

Content Creation tool needs to be more useful and flexible to use.

Combined the text and image generator function and redesigned the page to be more visible.

CC 1.png

Updated  'Content Creation' page

Discussion
Results

  • Incorporation of AI has shown to greatly expedite the user’s work progress as it is always anticipating and calculating the next best move for users.

  • Participants of the user testing have all expressed concurrence with the effectiveness of having AI to manage their schedules, tasks, and customer relationships. 

  • Despite its usefulness, many users still express distrust with AI.

  • During the user testing, most participants chose to complete the tasks manually instead of asking Atlas for help. 

Limitations

1. Lack of user testing sessions.
Due to the timeline of the project, only one round of user testing was accommodated after the low-fi prototype was designed. After which, the high-fi prototype was not tested with any users. For the scale of this project, it is imperative to have more testing sessions so that all changes made can be better justified. 

2. Heuristics evaluation was conducted by myself.
Ideally, a heuristics evaluation should be conducted with a group of usability experts so that consensus on issues can be raised. Since I was the only one who did the audit, there may be more usability issues that I have failed to identify. Likewise, there may also be issues I identified that may not be significant.

3. Functional limitations
Chats syncing feature across social media will require strict data governance and collaboration with the different business entities (Meta and Bytedance). Hence, without such assurances, users will question the safety of their privacy and will not be willing to use the app. Additionally, the quality of Atlas’ AI suggestions and content generated is dependent on how well the AI is written. Future developers can consider collaborating with existing AI technologies in the open market as they have a longer history of data to ensure better generations.

bottom of page