AI UX Patterns

Abstracted design solutions to review the core elements & interactions in each pattern. To reflect on the best-practice fit for your context when designing in the new AI paradigm.

Layout

Raw Text Input

The most basic and straight forward input, raw text directly prompting the AI.

Image Input

Adding an image accompanying a supporting text prompt.

Inline Help

Subtle help information presented inline with other UI elements.

Voice Input

Similar to raw text but more natural using voice input.

Inline Suggestions

AI powered suggestions to complete prompts while being authored.

Prompt Quality Feedback

Providing indication of a prompts quality or level of details while it's being authored.

Structured Prompt

Providing the required structure & detail required for the prompt. Scaffolding the users input.

Paginated Prompt

Like the structured prompt accept paginated to create a focused workflow.

Editing Assistance

Pre prompt submission AI assisted prompt editing.

Configurable Controls

Highly abstracted prompt authoring using configured options & controls.

Cloze Passage

Prompt variables presented in a conversation cloze passage style.

Reference Material

Adding specific files for the AI to reference to more tightly scope your resuklts.

Prompt Templates

Allowing prompts to be saved & recalled, as well as a library of pre written prompts.

Model Selection

Providing user control over the exact LLM modal they're using.

Thread Options

Providing users control over threaded or a fresh generation each time.

Thread History

Review & access a history of your various thread activities.

Generation Tokens

Providing an allocated amount of generation tokens to spend.

Result Options

Providing users useful options to interact with generated results.

Result Actions

Providing contextually relevant options per generated result.

Show Citations

Providing clear sources used to generate the results.

Timeline

Displaying results in a timeline when contextually relevant.

Object Oriented

Displaying results using their object categories to expose hierarchical insights.

Full Result Regeneration

A complete regeneration of the AI result.

Partial Regeneration

Focused regeneration of specific sections of an AI result.

AI Agent Initial Command

Defining a clear initial command for an AI agent to execute.

Agent Action Review & Confirm

Reviewing an agents understanding & providing confirmation.