Designing for the Moment You Switch Back
Most software remembers where information lives. It fails to remember where the work stands.
Think about using an AI agent. You give it a task. You switch to other work. You return ninety seconds later. The agent says "Done."
Now you face a problem. Did it just make a draft? Did it change real data? What evidence supports this? Is a human decision required?
To answer these, you must hunt through chat logs, notifications, and tabs. The system kept the activity but lost the context. This gap is why we need Workstream Continuity Design (WCD).
WCD ensures the interface remembers the state of work every time you switch.
Traditional design assumes you are the only worker. Modern design must account for software working while you are away. You might supervise several things at once:
- An agent researching a topic
- An agent writing code
- A customer reply
- A blocked deployment
The main task becomes the "switch-in." You enter a workstream, understand the state, make a decision, and leave. If you have to rebuild context every time, you lose time and focus.
WCD uses a specific grammar to help you orient fast: • GOAL: The outcome we want. • ATTN: Why you need to look now. • STATE: The current condition. • DELTA: What actually changed. • ACTORS: Who owns it and who acts next. • AUTH: What is permitted. • EVIDENCE: What supports the state. • EFFECT: The scope and risk. • NEXT: The safest next step.
This is more useful than a simple status light. It is faster than reading a long chat history.
Success in WCD is not about how fast users click. It is about how fast they understand. We measure this through:
- Time to Orientation: How fast you see if you need to act.
- Time to Decision Readiness: How fast you understand the change.
- False-ready rate: How often you think work is done when it is not.
The goal is a coherent model where software supports human oversight.
Read the full research edition on Zenodo.
Optional learning community: https://t.me/GyaanSetuAi