AI Product Design
I've worked on early AI concepts and shipped AI-enabled products. I bring practical AI product judgment: where AI adds real value, where it creates risk, and how to make it usable.
- AI product fit: identify when AI meaningfully improves a workflow vs. when it feels forced
- Trust and confidence: design for uncertainty, errors, confidence signals, and user control
- Hands-on fluency: use AI daily for coding, design, video, avatars, writing, strategy, and prototyping
1. Conversational app creation
Situation:
UiPath Apps was a low-code app builder where users could design applications by dragging controls onto a canvas, configuring rules, and connecting apps to automations and data. This worked for RPA and citizen developers, but starting from a blank canvas still created friction for business users who wanted to move from idea to working app faster.
Task:
Propose a new AI-assisted creation path that would complement the existing drag-and-drop designer. Instead of requiring users to start manually, the feature would let them describe an app in natural language, upload a PDF or image, or use an existing data entity to generate a working starting point—then refine it in the existing visual designer.
Action:
In 2020, before ChatGPT and before AI-native development tools like Cursor became mainstream, I proposed this direction to leadership, secured product and engineering support, and led the design effort. I shaped the shift from canvas-first creation to AI-assisted app generation, including prompt input, example guidance, attachment-based creation, generated screens, review flows, and editing paths back into the visual app designer.
Result:
The work established an early foundation for AI-assisted app building at UiPath. The concept later shipped as part of Autopilot in Apps, which helps users generate apps from natural language prompts, PDFs or images, and existing data entities, while preserving the existing drag-and-drop designer for editing and refinement.
Overall context
-
UiPath acquired a low-code app built by a single developer in South Africa. The technology was strong, but usability feedback was negative. We began redesigning it. The drag-and-drop designer video below captures interactions I proposed—many of them shipped.
Drag-and-drop app designer -
I proposed building an AI assistant into the app builder. The video below captures the essence of that proposal—and, about three years later, what the engineering prototype looked like.
AI-driven app building: concept (2020) and working prototype (2023) -
I left UiPath before the full rollout, but the AI-assisted path later shipped as Autopilot in Apps, now part of the UiPath suite. You can read about it in the UiPath Apps documentation .
2. Smart copy/paste app (“Clipboard AI”)
Situation:
Enterprise users were still wasting time manually moving data from documents, receipts, IDs, and other sources into business systems.
Task:
As the design manager on the project, my role was to guide the product experience, make the AI capability easy to understand, and help communicate the product vision clearly.
Action:
I helped shape the core experience: extract structured data from a document, show what the system understood, and paste it into the right fields in the target workflow. I reviewed and directed design work, refined key mockups, clarified the flow, and helped make a complex AI interaction feel simple, useful, and trustworthy.
Result:
UiPath Clipboard AI was recognized by TIME as one of the Best Inventions of 2023 . As the design manager on the project, I oversaw the design direction, helped shape the experience, and refined key mockups and product storytelling assets. The project became a strong example of my ability to lead design for early AI products, translate technical capability into clear user experience, and communicate an emerging product idea in a way people immediately understand.
Other AI-assisted product experiences
-
UiPath Autopilot Concept 1
-
Document Understanding
-
UiPath Autopilot
-
Coaching App