top of page

Context

Cognum, an AI-based Startup from New York, gave me the following UX challenge:

“This screen should be designed for users to easily work with highly independent AI employees.”

At the same time, I received another challenge more focused on UI (available here).
This UX challenge was completed in 12 hours and 41 minutes and you can check the events in chronological order below:

▶️ AFTER-TESTING.png

Final version. Access the navigable prototype here.

Adam, the user who is an AI.

The Artificial Intelligence called Adam is capable of creating and carrying out tasks with a high level of independence, in other words, it is a virtual employee.

So my first challenge was to map the briefing guidelines into two categories: Adam's needs and the user's needs.

Adam's Needs

  • Should understand the project's goals, plan out tasks in a timeline.

  • Work with minimal guidance from the user.

  • Be able to suggest and carry out tasks on its own.

  • Be able to update the project's status as it moves along.

User's Needs

  • Can create and oversee projects.

  • Be able to modify projects and tasks as they see fit.

  • Be able to see how the AI employee is progressing with tasks and projects.

  • Have various ways to manage and interact with projects, tasks, and the AI employee.

Timeline - The link between machine and human being.

The timeline is a visual tool that allows you to quickly locate past, present, and future events. To ensure that the user can easily understand all of Adam's completed, ongoing, and upcoming tasks, I have implemented this strategy.

sketch-Daniel-Costa 1.png

Meu processo de Design envolve muitos rascunhos, mesmo que a parte de pesquisa não esteja concluída. Isso me ajuda a identificar pontos cegos e fazer novos questionamentos.

The AI-Employee does it, but I'm in charge.

This project aimed to reduce the need for prompts entered by the user and increase the autonomy of the AI.

Therefore, the prompt that defines the objective of each project - where the user wants to go - would be the most important.
From there, the AI ​​itself would define the tasks necessary to achieve this objective and the sub-tasks involved.

INFORMATION ARCHITECTURE (2).png

The information architecture was created based on the project’s briefing.

ADAM’S FLOW (HYPOTHESIS) (1).png

Despite this great Adam’s autonomy, the user shall be able to intervene in the process by:
 

  • Editing project objective.

  • Editing or delete tasks.

  • Changing the priority level of tasks.

image 158.png

What should a good AI interface provide?

Desk research helped me understand key points to improve the usability of an AI interface.
And of course, I had a very interesting conversation with 'someone' who understands the subject very well - The GPT Chat.

image 10.png

Chat GPT presented me with 12 topics about making a good AI interface.

Will the relationship between a person and a machine work? True time.

I tested an initial version with a layperson unfamiliar with AI.
The test aimed to determine if anyone could
intuitively understand the screen with minimal context.

Frame 159 1.png

I faced issues with the testing platform (Maze), but I was able to proceed via video call. But since the person's audio wasn't working, she sent me her thoughts and sensations via audio messages on another social app.

Audio chat in a social app.

The test resulted in 3 insights that we can summarize with:

"Simplify the screen so that the primary elements are more noticeable and used".

(If you want to see the complete documentation for this test, click here)

Final Design Patterns.

Frame 175.png

Unique Codes that represents each type of task, like: ‘ORG’ - Organization.

Frame 176.png

Loaders in the current task and its done percentage.

Frame 177.png

Add, view, or edit tasks by clicking on the timeline.

Frame 178.png

Discuss a task with Adam.

Frame 179.png

Minimized elements that increase the prominence of primary components.

Frame 180.png

Main visual hierarchy for the primary button 'Add a task'.

Frame 181.png

Pause task immediately.

Frame 182.png

Component (moving line) that indicates the ongoing task.

Frame 183.png

Header (with general actions) differentiated with shadow.

Presentation to Stakeholders and the final result.

The stakeholders were greatly satisfied with my unwavering dedication and determination to tackle both challenges. Throughout the process, I made sure to delve deeply into every aspect of the task at hand, analyzing all the available information to extract the best possible result. 

My commitment to this approach is the key to achieving success, and I will continue to employ it in all my future endeavors.

[!] I have summarized the main points in this study. For more detailed information, please refer to My Figma.

Apprenticeship.

It's possible to create and validate good ideas for Digital Experience Design even with limited resources and time.

bottom of page