Pattern 13
Partial is not complete
Separate interrupted output from missing source or conversation context.
Concrete takeaway
Label an interrupted generation and offer continuation or retry.
Choose an approach
These are editorial decision conditions to validate. Some alternatives are successive states or can be combined; they are not always exclusive choices.
| Alternative | Use when | Tradeoff or requirement |
|---|---|---|
| A · Interrupted answer | Generation stopped before completion | Continuation may duplicate text unless recovery is anchored |
| B · Missing source | An earlier source is outside active context | Ask for it rather than inventing recall |
| C · Context review | A summary is used as current context | The person must be able to inspect and correct it |
The problem
A stopped answer may look finished. A long visible conversation may also imply that the model still has access to every earlier message.
The pattern
Label an interrupted generation and offer continuation or retry. Identify relevant unavailable context and ask for the missing information. Keep confirmed memories, source text and conversation summaries distinct.
Compare the alternatives
Separate interrupted output from missing source or conversation context.
A · Interrupted answer
An unfinished result has an explicit state.
B · Missing source
The assistant does not invent access to a document.
C · Context review
A summary can be checked rather than treated as perfect memory.
Original wireframe proposals · No live controls · Grey rows represent schematic text, not loading states · Labelled placeholders are not Scripture quotations
Download this wireframe as SVGFailure modes
Calling a partial result final; invented recall; a generated summary silently replacing verified text or user-approved memory.
What to validate
Interrupt generation and remove an earlier source from active context. Does recovery avoid duplication, disclose gaps and retain the user’s original intent?
Research-informed design proposal. These visual alternatives have not been tested with users in this atlas. Source findings, documented behaviour and this proposed adaptation are different kinds of evidence.
Sources and adaptation
This card is an original design synthesis. The following sources inform its content distinctions, interaction approach or review requirements; they do not validate the whole pattern.
- Interrupt and ResumePattern detail reviewed · Backend recovery must match the UI promise.
- Design Patterns for AI InterfacesPublic syllabus reviewed on 5 October 2026 · The course videos were not reviewed. Syllabus content is not treated as tested guidance.
- AI Interaction AtlasOfficial repository and taxonomy reviewed · A taxonomy, not a prescriptive UI framework. The official repository is available at https://github.com/quietloudlab/ai-interaction-atlas under Apache 2.0.
Common scenarios and flows
Use this pattern within a complete journey: define the entry, preserve relevant state, plan recovery and name the endpoint. Explore the scenario library.
Design principles in this decision
Editorial application of Missional by Design. These values frame review questions; they do not validate a pattern’s effectiveness.
- Hospitality — Who can use this approach in their actual context, and what complete alternative exists for someone who cannot?
- Honesty — Could this interface lead a person to infer more authority, certainty, privacy or human support than the service provides?
- Costliness — Who pays for this convenience through time, labour, privacy, risk or exclusion, and can the organisation support its promise?
Compare alternatives using the task and evidence above. Record competing needs instead of treating a principle as an automatic verdict.