Give the workflow a visible structure.
I organized the workbench into Template Definition, Examples, and Content Generation. Separating reusable setup from generation gives people a recognizable place to return to each part of the process.
LIONBRIDGE / CONTENT REMIX
Making multilingual content creation
easier to manage.
I owned end-to-end UX and UI design for Content Remix, bringing reusable briefs, AI generation, and content review into one workbench.
CONTENT REMIXDEFINE · GENERATE · REVIEW
01 / THE CHALLENGE
The prompt is only one part of the product.
Creating content with AI means making several connected decisions: what source material to use, what to produce, which languages to support, and how to guide the result. A useful product has to make those choices understandable and the outputs manageable.
I owned UX and UI across Content Remix and took on product-management responsibilities for the application. My work connected research and journey mapping with detailed interfaces and designs supporting development planning.
DESIGN QUESTION
How can one workbench support a clear first generation—and the structure people need to repeat the work?
02 / BUILDING THE FOUNDATION
Understand the work around generation.
I studied interface patterns in tools including Copy.ai, Jasper, and Grammarly: navigation, prompt guidance, settings, and how generated content was presented. I also mapped user journeys, roles, permissions, and subscription scenarios.
That work helped connect the creation experience to the account and customer structures around it. The design needed to support a reusable workflow, as well as an individual request.
01
Template definition
Make the brief explicit.
02
Examples
Give the setup context.
03
Content generation
Produce and review outputs.
Decisions that shaped the experience
Intent → Structure → ControlMake the content brief explicit.
The template asks for a name, project goal, input fields, and output types. Optional persona, tone, and specific instructions sit in an additional-information section, keeping core requirements distinct from extra guidance.
Make the language choice a real choice.
Generating separately in each target language and translating from one source language are different strategies. I gave them explicit controls so the choice is visible in the setup.
Make every output reviewable.
Channel-specific results have their own copy and edit actions, with word and character indicators. Language tabs organize the results so people can review one output without navigating a single long response.
Design for repeated work.
Reusable templates, spreadsheet entry points, generation cards, and expand/collapse controls support work beyond a single request. Manual and batch entry points offer distinct ways into content creation.
Keep source inspection inside the workbench.
For image-based use cases, previews open within the generation context. This keeps the source available while someone works with the resulting product description, travel content, or transcription.
04 / DELIVERY & REFLECTION
Connecting the product direction to the details.
I produced intermediate designs for planning and detailed interfaces for development. The work connected template configuration, generation, multilingual output, and source inspection into a consistent experience.
This project reflects how I approach AI products: designing the work around the model, including how people express intent, understand options, repeat a task, and review what comes back.
COMPANY RECOGNITION / 2025
Training Industry’s Top 20 AI Content Creation and Authoring Tools Companies
Lionbridge was selected for the 2025 list, with Content Remix highlighted in the company announcement.
Read Lionbridge’s announcement