Skip to content

Draft v0.1

Display Results

Lifecycle Stage 2
This article is part of the AI Interaction Lifecycle framework
Express Intent → Display Results → Refine Output → Take Action → Execute & Monitor → Summary

Overview

Once an AI system has interpreted a user's intent, it must communicate the outcome in a way that is understandable, trustworthy, and actionable.

Display Results focuses on the patterns that help users evaluate AI-generated outputs, understand how conclusions were reached, verify supporting information, and determine whether additional refinement or action is needed.

As AI systems become more capable, displaying a result is no longer limited to showing an answer. Results may include generated content, recommendations, reasoning artifacts, sources, plans, actions, or execution outcomes. Effective result presentation helps users build confidence in the output without overwhelming them with unnecessary complexity.

Whether the experience is informational, assistive, or agentic, the goal of this stage is to help users understand what the AI produced and decide what to do next.

Patterns

Explainability Layers

Explainability Layers provide multiple levels of explanation that help users understand how an AI-generated result was produced. Rather than overwhelming users with technical details, information is progressively disclosed based on their needs and level of interest.

These layers may reveal:

  • Data sources used to generate the result
  • Supporting evidence and references
  • Decision pathways or reasoning summaries
  • Confidence indicators
  • Actions taken on the user's behalf

By presenting explanations at varying levels of detail, users can better understand outputs and determine whether additional verification is required.

Human-Verified vs. AI-Generated

The Human-Verified vs. AI-Generated pattern clearly distinguishes between information that has been reviewed or approved by people and information created solely by AI.

This distinction helps users:

  • Understand the origin of content
  • Assess the appropriate level of trust
  • Make informed decisions about relying on outputs
  • Meet organizational and regulatory requirements

Clear attribution increases transparency and helps establish appropriate expectations for accuracy and accountability.

Stream of Thought

A Stream of Thought provides a visible trace of how an AI system navigated from input to output.

Depending on the context, this may include:

  • Tasks the system planned to perform
  • Tools or services it invoked
  • Information sources consulted
  • Actions taken on the user's behalf
  • Reasoning checkpoints and validation steps

By making portions of the AI's process visible, users gain additional context that can help them understand, evaluate, and trust the result.

Footprints

Footprints are the visible and machine-readable traces that show where and how AI participated in creating, modifying, or influencing content and decisions.

Examples of footprints include:

  • AI-generated content indicators
  • Edit histories
  • Activity logs
  • Provenance metadata
  • Audit records

Footprints help users understand how outcomes were produced, support governance and auditing efforts, and assist organizations in meeting compliance and regulatory requirements.

Branching

Branching helps users explore alternative paths, consequences, and outcomes by visualizing relationships between decisions and their potential effects.

This pattern can be used to:

  • Compare different approaches
  • Understand cause-and-effect relationships
  • Evaluate tradeoffs
  • Explore scenarios and decision paths
  • Assess potential risks and opportunities

By making alternatives visible, branching supports more informed decision making and reduces reliance on a single AI-generated recommendation.

Citations

Citations help users understand where AI-generated information originated. They connect generated outputs back to source materials such as documents, transcripts, web pages, databases, or enterprise knowledge sources.

The primary goals of citations are to:

  • Increase transparency
  • Support verification
  • Provide supporting evidence
  • Improve confidence in outputs
  • Enable deeper exploration of source material

Citation Variations

Inline Highlights

Best suited for attached content such as PDFs or documents.

Example: Adobe Acrobat highlights source passages directly in the summary experience, allowing users to jump to the original content.

Direct Quotations

Best suited for transcripts, meeting notes, or long-form content.

Example: Granola presents the specific transcript excerpts that support key takeaways and summaries.

Multi-Source References

Best suited for search, research, and aggregated responses.

Example: Perplexity displays multiple citations with supporting metadata, helping users quickly evaluate source relevance.

Best suited for simple verification scenarios.

Example: Copy.ai presents source URLs alongside generated content, prioritizing transparency and traceability.

Caveats

Caveats communicate that AI-generated outputs may be incomplete, incorrect, outdated, uncertain, or biased.

Common caveats include:

  • Confidence indicators
  • Disclaimer messaging
  • Risk notifications
  • Known limitations
  • Data freshness warnings

While caveats are commonly used across AI experiences, they should not be relied upon as the sole mechanism for encouraging critical thinking. Effective AI experiences combine caveats with transparency, verification tools, citations, and user controls to support informed decision making.

Data Privacy Controls

Data Privacy Controls give users visibility into and control over how their information is collected, stored, shared, and used by AI systems.

These controls may allow users to:

  • Manage stored memories and preferences
  • Control data retention
  • Review permissions
  • Opt in or out of personalization features
  • Understand how data contributes to model behavior
  • Meet privacy and regulatory requirements

Providing clear privacy controls strengthens trust and helps users feel confident that their information is being handled responsibly.

Principles Applied

This stage primarily aligns with the following AI UX principles:

  • Transparency – Users should understand how outputs were generated and where information originated.
  • Trust & Confidence – Users should be able to appropriately evaluate and rely on AI-generated results.
  • Control – Users should be able to verify, challenge, and further investigate outputs.

Next Stage: Refine Output

Once users understand the AI's response, they often need to improve, redirect, or further develop the outcome.

The next stage in the lifecycle, Refine Output, focuses on the patterns that help users iteratively improve AI-generated results through feedback, scoping, contextual understanding, and refinement actions.