# How should users and AI agents share control of a task?

Let users co-plan and assign steps, interleave planning with doing, and hand control back on risky steps. Early studies show this steers agents without making them hard to use.

Evidence: Limited (Early signal, treat it as a hypothesis) · LLM fit: Medium (Tested in lab or early LLM studies, likely carries over) (https://aiuxevidence.com/how-we-rate)
Two builder-run system studies with small samples and simulated users.

## Design it like this

Make the plan editable, let steps be assigned to the user or agent, and support easy handover.

- Show the plan before execution and let users edit it. [from the study]
- Let users assign individual steps to themselves or the agent. [from the study]
- Allow the plan to change after partial execution. [from the study]
- Ask for help sparingly; it was needed in a minority of tasks. [from the study]

## What they found

### Feng et al. (2026), system evaluation

After a formative study with 9 researchers, the authors built Cocoa, a notebook-style agent where plan steps can be assigned to the user or the agent and planning and execution interleave. A lab study (16 people) found it enabled steering without sacrificing ease of use; a week-long field deployment (7 people) showed it used on real research tasks. Builder-run evaluation.

> “Cocoa enabled steerability without sacrificing ease-of-use” (Abstract)

### Mozannar et al. (2025), system evaluation

A GPT-4o web agent with co-planning, action approval for irreversible steps, and handover to the user. On 162 GAIA validation tasks, a simulated user with side information raised task completion from 30.3% (autonomous) to 51.9%, while the agent asked for help in only 10% of tasks. A qualitative study with 12 users explored the features; this is a system paper by its builders, not an independent test.

> “Magentic-UI implements an action guard system to ensure irreversible or potentially harmful agent actions are reviewed by the human user.” (Figure 8 caption)

## Trade-offs and when it doesn't apply

- Both are builder-run studies.
- Small samples (16 lab, 7 field, 12 users).
- Simulated users drove the Magentic-UI gains.

## What it looks like

- Do: An agent plan with checkboxes: “You do step 3”, “Agent does step 4”.
- Don't: A black-box agent that runs to completion with no checkpoints.

## How to check it in your product

Track how often users edit plans or take over, and whether those tasks succeed more often.

## Open questions

- Which steps should default to the user?
- How much plan detail helps before it becomes noise?
- Do independent evaluations replicate these benefits?

## FAQ

### Do humans in the loop improve agent results?

In Magentic-UI's evaluation, a simulated user raised task completion from 30.3% to 51.9% on 162 tasks.

### Does co-planning make agents harder to use?

Cocoa's lab study (16 people) found it enabled steering without sacrificing ease of use.

## Related questions

- [When should an AI agent ask for confirmation before acting?](https://aiuxevidence.com/e/ask-before-acting)
- [Should users be able to edit AI-generated output?](https://aiuxevidence.com/e/let-users-edit)
- [Does relying on AI reduce users' critical thinking?](https://aiuxevidence.com/e/deskilling)

## Sources

1. K. J. Kevin Feng, Kevin Pu, Matt Latzke, et al. (2026). Cocoa: Co-Planning and Co-Execution with AI Agents. CHI 2026 (arXiv:2412.10999v4). https://doi.org/10.1145/3772318.3791673
2. Hussein Mozannar, Gagan Bansal, Cheng Tan, Adam Fourney, et al. (2025). Magentic-UI: Towards Human-in-the-loop Agentic Systems. arXiv:2507.22358 (Microsoft Research, 2025). https://arxiv.org/abs/2507.22358

Canonical: https://aiuxevidence.com/e/agent-handoff