Live Plan Surface
The plan stays editable during execution — both human and AI can grab any step
Most plan UIs (Editable Plan, Intent Preview) are pre-execution: the user reviews the plan, edits it, and then hands off to the agent. Co-planning is different — the plan stays a live, editable surface *during* execution. Either side, human or agent, can take ownership of the next step, insert a new step, reorder, or hand back. The plan is the workspace, not just a contract. When the user grabs a step, the agent pauses on that branch. When the agent grabs the next step, it leaves a receipt the user can inspect. Adapted from Feng, Pu, Latzke, August, Siangliulue, Bragg, Weld, Zhang & Chang ("Cocoa: Co-Planning and Co-Execution with AI Agents", Allen AI + University of Washington, Dec 2024 / refined Feb 2026, CHI 2026), who showed that co-planning beats both full automation and full manual on outcome quality and user trust.
Framing
The problem
Plans go stale the moment execution starts — the user cannot intervene mid-run without aborting and re-prompting from scratch.
The pattern
Keep the plan editable during execution; let either side take ownership of any step at any time, with explicit ownership semantics and per-step receipts.
Why chat breaks here
Chat moves the plan into the transcript history once execution starts — edits would have to be re-prompted from scratch, breaking the agent's state.
Risks
Without clear ownership semantics, human and agent can clash on the same step; without receipts, completed steps cannot be re-inspected after the fact.
Avoid when
The plan is short or low-stakes enough that pre-execution review (Editable Plan) is sufficient and live ownership transfer adds complexity for no gain.
Use when
Multi-step agent runs where the user wants to step in mid-flow without aborting — research, multi-document production, code reviews where intervention has to be cheap and surgical.
DOPE evaluation
- Directability
- Drag a step to a different owner, insert a new step inline, skip or reorder mid-run — the agent re-plans around the change
- Observability
- Each step shows its current owner (AI / human / unassigned) and the receipt for completed steps — what was done, by whom, with what evidence
- Predictability
- Ownership is explicit per step; the agent never silently takes a step the user marked human-only, and vice versa
- Explainability
- Each ownership change is logged with reason; each completed step links to the receipt the executor produced
In the wild
- Claude Code · Live task list (Anthropic) — The agent maintains a visible task list (the Tasks API, formerly TodoWrite) that updates as it executes — steps are added, reordered, and checked off mid-run, and the user can redirect between them. The plan and the execution share one surface.
- Cursor · Plan Mode (Cursor) — Plan Mode ships an interactive editor for modifying plans inline, and Composer keeps the plan open to refinement while the agent executes — edit a step mid-run and the agent re-plans around it.
- GitHub · Copilot app Canvas (GitHub) — The canvas is explicitly bidirectional: the agent updates it as it works, and the user can edit, reorder, approve, or redirect work on that same surface (GA June 2026). The best-verified production match for a co-owned plan.
- Manus · Task list (Manus AI) — The plan updates on the fly as the agent works; the user can jump in mid-run or take control of a step, and replay what each step did. Take-over is real, though per-step ownership is not as explicit as the pattern asks.
FAQ
When should I use the Live Plan Surface pattern?
Multi-step agent runs where the user wants to step in mid-flow without aborting — research, multi-document production, code reviews where intervention has to be cheap and surgical.
When should I avoid the Live Plan Surface pattern?
The plan is short or low-stakes enough that pre-execution review (Editable Plan) is sufficient and live ownership transfer adds complexity for no gain.
What problem does Live Plan Surface solve?
Plans go stale the moment execution starts — the user cannot intervene mid-run without aborting and re-prompting from scratch.
Why is chat the wrong fit for this?
Chat moves the plan into the transcript history once execution starts — edits would have to be re-prompted from scratch, breaking the agent's state.
Related patterns
- Extends: Editable Plan — Editable Plan is pre-execution review — you negotiate the plan, then hand off. Live Plan Surface extends editability *through* execution, so intervention does not require abort + restart.
- Alternative to: Intent Preview — Intent Preview is a snapshot before kickoff; Live Plan Surface is a continuous-edit surface during the run. Different time horizons for the same scope-visibility concern.
- Often paired with: Mid-Stream Steering — Mid-Stream Steering nudges the generation token-stream; Live Plan Surface offers step-level ownership transfer. Two granularities of the same "intervene without aborting" move.