ContidaAI Web-based product design
A constraint-aware AI workspace designed for professional work that spans multiple tasks and evolving context.
About Contida AI
Contida AI is a project-based AI workspace built for professionals working across coding, research, analysis, and drafting. Unlike generic chat tools, Contida preserves project-specific context such as constraints, standards, and intent across multiple tasks within the same workspace.
The system is designed for environments where accuracy, alignment, and continuity matter more than speed alone.
Project type
Product Design
Role
UIUX Designer
Duration
2 weeks
The Problem
Professional users increasingly rely on AI for complex work, but most AI tools assume each interaction is isolated. Context is forgotten, constraints are lost, and users are forced to repeatedly restate standards, assumptions, and intent.
The issue is not that AI is “wrong,” but that current tools are not designed to carry responsibility across a project lifecycle.
Design Constraints & Intent
AI output cannot be assumed correct by default
Professional work requires continuity, not just answers
Verification is part of the workflow, not a separate step
The system must support uncertainty without overwhelming the user
A project in motion
When users open Contida, they are not dropped into a blank conversation. They are presented with a list of projects — ongoing and completed — each representing a distinct body of work with its own context, constraints, and history.
From this screen, users can:
Review existing projects
Reopen work exactly where it was left
Start a new project with a single action
To begin new work, the user creates a project rather than a conversation.
The New Project screen opens directly into a chat interface, accompanied by a brief system message that frames the interaction as an ongoing workspace rather than a one-off prompt.
Before the research begins, the user defines how the project should think.
Project Directives act as a persistent layer of intent — setting boundaries around tone, sources, assumptions, and acceptable uncertainty.
Directives are organized by domain (Research, Code, Design, Analytics), allowing the same project to support different types of work without losing context.
This setup explains why some outputs later carry review anchors.
Showing explicit signaling of uncertainty rather than confident guesses.
With project directives set, the user begins an ongoing research conversation.
The interface remains familiar — a continuous dialogue.
Some parts of an output carry more uncertainty than others.
Rather than interrupting the flow, Contida marks these moments subtly. Inline anchors & review signals
When the user hovers over an anchor:
The relevant paragraph is softly highlighted
A tooltip explains why the section was flagged
As the research progresses, reviewed outputs can be saved into documents — turning transient conversation into structured work.
The user needs a way to retain meaningful artifacts without preserving entire conversations.
The Library collects documents and media explicitly saved from projects and sidebars. Items are organized by project, not chronology.
What This Project Demonstrates
Contida AI explores how an AI workspace can maintain project continuity across complex professional tasks.
Throughout the walkthrough, the system centers on a single idea: work does not happen in isolated prompts. Research, decisions, references, and outputs accumulate over time, and losing that context creates friction, repetition, and cognitive overhead.
The outcome is a workspace concept that prioritizes continuity over novelty, and structure without rigidity — designed to support real, multi-session work rather than isolated interactions.
