AIX//Artificial Intelligence Experiences/AIX
[UX/UI/AIX]
Patterns//_
To consider the interaction contract between humans and machines./
AIX
Artificial Intelligence Experiences/AIX//
[UX/UI/AIX]
Patterns To consider the interaction contract between humans and machines./
UX/UI/AIX
Patterns
For the interaction contract between humans and machines.
UX/UI/AIX
Patterns
For the interaction contract between humans and machines.
UX/UI/AIX
Patterns
For the interaction contract between humans and machines.
UX/UI/AIX Patterns are key for designers and developers building AI-first products. In fact, it’s crucial to consider the interaction contract between humans and machines.
UX/UI/AIX Patterns
for Designing
LARGE
Multimodal Models
of GenAI experiences.
UX/UI/AIX Patterns
for Designing
LARGE
Multimodal Models
of GenAI experiences.
UX/UI/AIX Patterns
for Designing
LARGE
Multimodal Models
of GenAI experiences.
20+ Next-gen
UX/UI/AIX
Patterns
engineered to power
human-centric design
in Generative AI—where AI
interface blends true intelligence.
20+ Next-gen
UX/UI/AIX Patterns
engineered to power
human-centric design
in Generative AI—where AI
interface blends true intelligence.
20+ Next-gen
UX/UI/AIX Patterns
engineered to power
human-centric design
in Generative AI—where AI
interface blends
true intelligence.
Whether you’re working with chat-bots, generative tools, or autonomous agents, thoughtful interface design can transform AI from a black box into a truly empowering experience.
GenAI or No GenAI
Intercom's FinAl captures user intent and provide detailed explanations compared to a structured button based chat bot interface or FAQs list.
GenAI or No GenAI
Intercom's FinAl captures user intent and provide detailed explanations compared to a structured button based chat bot interface or FAQs list.
Data Needs
Convert User Needs to Data Needs: Paper "Towards Accountability for Machine Learning Datasets" outlines best practises of Data development cycle in software engineering.
Data Needs
Convert User Needs to Data Needs: Paper "Towards Accountability for Machine Learning Datasets" outlines best practises of Data development cycle in software engineering.
Augment vs Automate
Al tools like Magenta Studio in Abelton gives creative controls to manipulate and create new music without taking over the process.
Augment vs Automate
Al tools like Magenta Studio in Abelton gives creative controls to manipulate and create new music without taking over the process.
Define Level of Automation
Github Co-Pilot is an example of Partial Automation. It suggests code based on context but user could edits, accepts or ignores suggestions.
Define Level of Automation
Github Co-Pilot is an example of Partial Automation. It suggests code based on context but user could edits, accepts or ignores suggestions.
Provide Data Sources
Google NotebookLM adds citations automatically. It links each answer directly to the part of the user's uploaded document where the information came from.
Provide Data Sources
Google NotebookLM adds citations automatically. It links each answer directly to the part of the user's uploaded document where the information came from.
GenAI Adoption
Progressive GenAI Adoption such as Adobe FlreFly progressively onboards user with basic to advance Al features.
GenAI Adoption
Progressive GenAI Adoption such as Adobe FlreFly progressively onboards user with basic to advance Al features.
Leverage Mental Models
Generative Expand in Photoshop builds on the familiar idea of expanding an image. Users simply extend the canvas and then prompt the tool to fill the new space.
Leverage Mental Models
Generative Expand in Photoshop builds on the familiar idea of expanding an image. Users simply extend the canvas and then prompt the tool to fill the new space.
Convey Product Limits
Convey Product Limits using AI. Claude Al assistants clearly state their knowledge cutoff dates or when they are unsure about an answer or when a question falls outside their training data.
Convey Product Limits
Convey Product Limits using AI. Claude Al assistants clearly state their knowledge cutoff dates or when they are unsure about an answer or when a question falls outside their training data.
Display Chain of Thought
Perplexity Al enhances transparency by displaying its processing steps helping users understand the thoughtful process behind the answers.
Display Chain of Thought
Perplexity Al enhances transparency by displaying its processing steps helping users understand the thoughtful process behind the answers.
Leverage Multiple Outputs
Google Gemini provides multiple options to help user explore, refine and make better decisions.
Leverage Multiple Outputs
Google Gemini provides multiple options to help user explore, refine and make better decisions.
Convey Model Confidence
Grammarly app uses verbal qualifiers like "likely" to the content it generated along with the user.
Convey Model Confidence
Grammarly app uses verbal qualifiers like "likely" to the content it generated along with the user.
Memory and Recall
Design for Memory and Recall. ChatGPT remembers key facts you've shared to personalize future chats. You can view, update, or delete these memories anytime.
Memory and Recall
Design for Memory and Recall. ChatGPT remembers key facts you've shared to personalize future chats. You can view, update, or delete these memories anytime.
Data Privacy & Controls
Communicate Data Privacy & Controls. Slack Al clearly communicates that customer data remains owned and controlled by the customer, and is not used to train Slack's or any third-party Al models.
Data Privacy & Controls
Communicate Data Privacy & Controls. Slack Al clearly communicates that customer data remains owned and controlled by the customer, and is not used to train Slack's or any third-party Al models.
Input Parameters
Provide Contextual Input Parameters. ElevenLabs lets you adjust several settings to design a custom voice.
Input Parameters
Provide Contextual Input Parameters. ElevenLabs lets you adjust several settings to design a custom voice.
Design for Co-Pilot
Design for Co-Pilot. Notion Al helps you draft, summarize, and edit content while you control the final version.
Design for Co-Pilot
Design for Co-Pilot. Notion Al helps you draft, summarize, and edit content while you control the final version.
Controls for Automation
Design User Controls for Automation. Users can enable or disable predictive text and auto-replies on Gmail app.
Controls for Automation
Design User Controls for Automation. Users can enable or disable predictive text and auto-replies on Gmail app.
User Input Error states
Design for User Input Error states. When given vague prompts like "What's the capital" the models often reply to clarify further context.
User Input Error states
Design for User Input Error states. When given vague prompts like "What's the capital" the models often reply to clarify further context.
Capture User Feedback
Design to Capture User Feedback. ChatGPT uses Reaction buttons (thumbs up or thumbs down), Comment boxes and Clarification prompts to collect user feedback.
Capture User Feedback
Design to Capture User Feedback. ChatGPT uses Reaction buttons (thumbs up or thumbs down), Comment boxes and Clarification prompts to collect user feedback.
AI Error States
Design for AI Error States. CitiBank financial fraud system displays the following error message because of False positives or false negatives.
AI Error States
Design for AI Error States. CitiBank financial fraud system displays the following error message because of False positives or false negatives.
Design for AI Guardrails
Design for AI Guardrails. Instagram uses hybrid approach of user feedback for Al to spot toxic content and misinformation and stop from it spreading.
Design for AI Guardrails
Design for AI Guardrails. Instagram uses hybrid approach of user feedback for Al to spot toxic content and misinformation and stop from it spreading.
Model Evaluations
Design for Model Evaluations. Amazon Bedrock, a managed Al service uses LLM-as-a-judge feature for model evaluation.
Model Evaluations
Design for Model Evaluations. Amazon Bedrock, a managed Al service uses LLM-as-a-judge feature for model evaluation.
.
1
3