OrchestratorAI
Bringing multiple AI models together in one interface.
An AI product workspace that coordinates specialist agents across Claude, ChatGPT and Gemini, giving users one place to define tasks, direct the work and review outputs.
Making multi-model collaboration understandable and controllable by connecting agent assignment, task context, sources, conversation and outputs in one workspace.
Contribution:
Led the project from product definition and research through concept exploration and UI design, taking the initial idea through to a working prototype.
Product strategy, UX direction, Interface & Prototype
OrchestratorAI
Design Showcase
The challenge
The founder had a vision for AI models working together. The challenge was defining how people would direct that work, follow its progress, and review the results.
Constraints
Limited evidence. Limited time.
Without a budget for user interviews, early research drew on Reddit and forums. Time-to-market pressure kept the focus on defining the core experience.
Approach
Explore, refine, and iterate. The focus was on shaping the core workflow into a prototype that could be tested and developed further.
What is the product?
OrchestratorAI is an innovative tool designed to help users solve complex problems by coordinating a team of specialized language models (LLMs). The orchestrator acts as a central system that intelligently assigns roles to different LLMs based on the user's input, creating a collaborative environment where each LLM contributes its expertise to generate high-quality outputs.
01
Supporting access and account recovery
Familiar sign-in options and a guided password reset flow give users a path into the workspace, with recovery instructions and support available when access is interrupted.
02
Structuring navigation around user intent
Navigation was organized around the user’s next decision: search for information, start a task, manage connected sources, configure or switch agents, and return to ongoing work without losing context.
03
Structuring task delegation to output review
The workspace connects the request, agent activity, generated artifacts, and version controls in one flow—giving users a clear way to follow progress, inspect different outputs, and revisit earlier work without losing context.
04
Project Framing & Context
Project setup turns an open-ended request into a defined job by establishing context, sources and how agents will be assigned.
05
Creating Agents & Roles
Agent roles, instructions and file access make the division of work visible before the task begins.
06
Bringing project knowledge into one place
The Database Hub brings connected datasets into a searchable collection, with descriptions and management controls that help users identify and maintain the sources supporting their AI tasks.
07
Turning AI outputs into usable deliverables
A dedicated artifact workspace keeps reports, diagrams, code and media alongside the conversation, with folders, version history and sharing controls to support review and use beyond the chat.
08
Bringing specialist agents together around one create artifacts
The concept brings agents with different roles into a shared task, connecting research, writing and review to develop a report while keeping user guidance and the resulting output in one workspace.