Vydeo
Designing a chat-based AI motion design experience that helps users turn ideas into polished animated content without needing to know how to animate.
Motion design explains ideas clearly. But making it has always required knowing how to edit, not just what to say.
Motion design is one of the most persuasive forms of communication a product team can produce. A well-crafted motion graphic can communicate a feature, a workflow, or a product value proposition in 15 seconds in a way that a paragraph can't match in 150 words. It gives information rhythm, hierarchy, and emotional weight.
For most people who need motion design: founders explaining their product, marketers launching a campaign, product teams announcing a feature, the bottleneck isn't creativity. It's execution. Traditional motion design tools ask you to think in timelines, keyframes, layers, animation curves, transitions, and export settings. That's the wrong starting point for someone who just needs to communicate an idea.
The design challenge this creates is more interesting than it first appears. Building a text-to-video generator is a generation problem. Building a tool where users feel confident directing the output: describing what they want, understanding what the AI understood, refining toward the right result is a product design problem. That's the work Vydeo needed done.
The real design problem was not just generation. It was helping users feel confident directing the output.
Vydeo is not a video editor. It's a motion design conversation.
Vydeo is an AI motion design platform that lets users create AI-powered videos and motion graphics by chatting with AI. Users don't manage a timeline or arrange layers, they describe what they want to create and the AI interprets that creative direction into motion graphic output.
This changes the UX problem fundamentally. Instead of designing a traditional editor with dozens of controls, the experience needs to support a conversational creation flow where each step builds from what the user said.
User describes an idea
The entry point is a natural language description of what the user wants to create, not a settings panel or a blank canvas.
AI interprets the creative direction
The product processes the intent, goal, audience, tone, format, visual style and builds a creative brief before generating anything.
Product generates motion design output
The AI creates motion graphic options based on the creative direction, not a single result, but multiple directions to compare and choose from.
User previews the result
A visual preview workspace gives users a real sense of the motion, timing, and composition before committing to a direction.
User refines through follow-up conversation or controls
Adjustments happen through natural follow-up: "Make it feel more premium," "Shorten this to 10 seconds," "Create a version for LinkedIn."
User exports, reuses, or iterates
The final moment creates momentum, not just one output, but a creative direction the user can remix, adapt, and reapply to future work.
Vydeo sits in a specific gap between AI video generators, motion tools, and creative assistants.
Most AI video tools generate footage. They're powerful, but they're still production-oriented aimed at users who already know what kind of output they're building toward. Vydeo's opportunity is more specific and more human: help users create structured motion graphic designs through conversation.
The product doesn't sit in one category cleanly. It draws from AI generation, motion design tooling, creative direction, and lightweight editing, but its actual value is in making those capabilities feel conversational and accessible, not just capable.
What users want to create
What Vydeo changes
Users know what they want to say. They don't always know how to tell the AI.
The most common user entry point isn't "I want to learn motion design", it's a specific communication need: announce a feature, explain a product, launch a campaign, make a concept feel real. The user has a clear goal and no clear path from here to there.
Chat-based creative tools surface a specific UX problem: the blank message box is not enough. Users need help knowing what to describe, how specific to be, and what the AI is capable of interpreting. Too vague and the output misses the intent. Too technical and the experience stops feeling conversational.
Questions users couldn't answer alone
What the interface needed to answer
Designing the user-facing experience for Vydeo's AI motion platform.
As Product Designer / Data Systems Collaborator, I led the end-to-end design of user-facing features for Vydeo's interactive platform. My work focused on making the experience clearer, more usable, and easier to evolve as the AI interface matured in translating product needs and evolving AI capabilities into interface patterns users could actually work with.
This wasn't just about making screens look polished. It was about figuring out where the workflow broke down for users, which AI states needed visible feedback, and how to structure a creative conversation that didn't require users to think like prompt engineers.
What I worked on
Focus areas
Category research clarified what users actually need from a chat-based motion tool.
To better understand the emerging AI motion design space, I reviewed similar products and how they framed workflows, templates, prompting, previewing, and export. The goal wasn't to copy category patterns, it was to identify what users had come to expect, and where those expectations created design obligations for Vydeo.
The research surfaced consistent patterns across the category. Users benefit from outcome-based starting points knowing what they can make before they know what to describe. Templates reduce creative anxiety because they give users something to react to rather than build from scratch. Prompting needs guidance because most users don't know how to give an AI useful creative direction without being taught. And generation needs preview loops because seeing partial motion is very different from seeing a finished result.
Category expectations surfaced
Where Vydeo's angle was different
Five interface problems that shaped every design decision.
Building a chat-based creative tool creates a specific set of design problems that don't exist in traditional editors. These five challenges defined the design work and each one had real consequences for how users could trust, use, and iterate with the product.
Designing beyond the chat box
A message input alone does not create a complete creative workflow. Users needed guidance before, during, and after generation, not just a place to type.
Reducing blank-prompt anxiety
Users needed help describing their idea without feeling like they had to write the perfect prompt. The system had to lower the bar for starting without lowering the quality of direction.
Making AI interpretation visible
The interface needed to show what the AI understood from the conversation, so users could trust the direction before spending time generating and see where to correct it if the AI missed something.
Supporting creative iteration
Motion design requires previewing, comparing, adjusting, regenerating, and saving directions. The experience had to support iteration without forcing users to restart from zero at every turn.
Balancing simplicity and control
The experience had to feel approachable for non-designers while still giving users meaningful control over style, timing, layout, and output without turning into the editor the product was trying to replace.
From editing timelines to directing motion through conversation.
The strategy was to treat chat as the beginning of the workflow, not the whole workflow. A user should be able to talk to the AI naturally, but the interface should provide structure around creative goal, format, audience, tone, visual style, preview, refinement, and export. Chat opens the door. The product carries users through it.
The ideal flow wasn't a single chat thread from input to output. It was a layered experience where conversation started the direction, a generated brief confirmed it, motion options expressed it, a preview workspace refined it, and an export moment made it reusable.
Choose or describe the goal
Start with what the user wants to make using a guided starting point or a natural description. The entry should reduce creative anxiety, not add to it.
Chat with AI to define the motion direction
The conversation builds creative intent, what the user wants to communicate, who it's for, how it should feel, and what format it needs to be in.
Confirm what the AI understood
A visible creative brief shows users the AI's interpretation before generation begins, giving them a checkpoint to correct the direction without spending time on the wrong output.
Generate motion graphic options
Multiple directions, not a single output. Users should be able to compare approaches and choose the one closest to their vision.
Preview and compare results
A real motion preview, not a thumbnail, it gives users enough context to evaluate what the AI created and decide what needs to change.
Refine through chat or controls
Adjustments happen conversationally: describe the change and the AI interprets it. Quick controls support direct editssuch as style, timing, color, and layout without requiring full regeneration.
Export, remix, or save as a reusable direction
The final moment creates momentum. Users export their output, but also save the creative direction for future work building a library of reusable motion styles rather than starting from scratch each time.
Six interface patterns that make chat-based motion design work.
Each of these patterns addresses a specific failure point in the chat-based creation flow. Together, they form the interface layer between what a user says and what they see on screen.
The entry experience opens with a simple question: "What are you trying to create?" No settings panel, no format picker, no blank canvas. Starter prompts: "Create a product launch video," "Turn this feature into a motion graphic," "Build a short explainer for social" give users a direction to react to rather than a blank field to fill. The goal was to make users feel like they're directing an outcome, not operating a tool.
Even with chat, users need structural help. Prompt helpers surfaced contextual fields: goal, audience, format, tone, duration, platform, brand style, motion intensity, not all at once, but progressively, as the conversation developed. The product built the creative direction from what users said, so they never needed to think about what a "good prompt" looked like.
Before generation, the product surfaced what it understood from the conversation: a lightweight creative brief showing objective, audience, visual tone, key message, format, and motion style. This gave users a checkpoint to verify and correct before spending time generating. Trust in AI products comes from visibility: users need to see what the AI understood before they can trust what it creates.
Results appeared as multiple motion directions: cards with a preview thumbnail, style label, duration, format, and key message. Each card offered clear next actions: Preview, Refine, Generate Variation, Save Direction. Creative work is comparative. Showing multiple directions forced users to actively choose rather than passively accept the first output, which produced better results and gave users a stronger sense of agency.
The preview workspace let users see real motion before committing to a direction. Refinement happened through follow-up chat, "Make it feel more premium," "Shorten this to 10 seconds," "Use a warmer color palette," "Create a version for LinkedIn" alongside quick controls for style, timing, text, layout, and brand updates. The key design challenge was making refinement feel continuous rather than forcing users to restart when they needed to adjust something.
The final state was designed to create momentum beyond a single output. Users could export their video, but also save the creative direction as a template, remix it for different formats, generate social cutdowns, or save the full direction to a project. A good AI creative tool shouldn't end at one export, it should help users build reusable creative systems they can return to.
Interface states designed for
Design system components built
What the design work made possible.
The design work helped clarify how Vydeo could make AI motion design feel more approachable through conversation, guided creative direction, preview states, and refinement loops. By shaping the experience around user intent instead of technical editing controls, the product could help users move from idea to motion with more confidence.
Reduced creative anxiety
Guided starting points and prompt helpers lowered the barrier to starting, users didn't need to know how to prompt before they could create.
Visible AI direction
The creative brief state gave users a checkpoint before generation making the AI's interpretation visible and correctable, not invisible and opaque.
Comparative creative review
Multiple direction cards gave users something to compare producing better final choices than a single generated result could.
Continuous iteration
Refinement through follow-up chat kept the workflow moving forward, users adjusted without restarting, which made the process feel less risky and more exploratory.
Reusable creative systems
Saved directions and templates meant the product accumulated value over time, not just one export, but a growing library of reusable creative directions.
What this project taught me about designing conversational creative tools.
Vydeo was a case study in designing creative confidence. The challenge wasn't only helping users generate motion graphics faster, it was helping them understand how to direct the AI, shape the result, and move from a rough idea to a usable animated asset without needing to become a motion designer.
The biggest design lesson was that chat-based AI products still need structure. A message box can start the experience, but users need feedback, checkpoints, previews, and clear next steps to trust what the AI is creating. Conversation without context is still just a blank prompt in a different shape.
The work also reinforced something I've seen across every AI product I've designed: the visible interface earns the trust, even when the underlying model does the work. The creative brief state, the preview workspace, the refinement loop weren't decorative. They were the difference between a product that felt like a black box and one that felt like a collaborator.