Skip to main content

AI glossary Models & Architecture

What is an agent agenda?

← All glossary terms
In short

An AI agent's live plan: the ordered next steps, open subtasks, blocked items and checkpoints it works through towards a goal.

Explained

An agent agenda is the working plan an AI agent keeps while it carries out a task. It lists the next actions, the subtasks still open, the items that are blocked waiting for data or approval and the steps already completed, and it is updated as tool results come in. The agenda comes from breaking the goal into steps and revising them as the agent learns more, an approach studied in research on planning before acting and on interleaving reasoning with actions. Showing the agenda in the user interface and in logs makes an agent's run inspectable and lets teams check it against the requirements and acceptance criteria.

What an agent agenda contains

  • Goal: the task as the agent understood it.
  • Next steps: the ordered actions it intends to take.
  • Open subtasks: parts of the work started but not finished.
  • Blocked items: steps waiting for data, a tool result or a human decision.
  • Completed steps: what has been done, with the evidence or tool output behind it.
  • Checkpoints: points where the agent must verify a result or ask for approval before continuing.

How agents build and update it

At the start, the agent breaks the goal into steps, which research calls planning or task decomposition. As it calls tools and reads the results, it marks steps as done, adds new ones it discovers and reorders the rest. Approaches such as plan-and-solve prompting make the model write a plan before executing it; approaches such as ReAct interleave short reasoning steps with actions so the plan adapts to what each action returns.

In long tasks, the agenda also acts as memory: it tells the agent, or a new session that picks up the work, where things stand.

Why exposing it matters

  • Transparency: users see what the agent intends to do before it does it.
  • Control: a person can correct the plan, remove a step or approve a checkpoint.
  • Debugging: when a run goes wrong, the agenda shows where the plan diverged.
  • Alignment with requirements: each step can be checked against the specification and acceptance criteria of the task.
  • Audit: a logged agenda is a readable record of how a result was reached.

Frequently asked questions

What is the difference between an agent agenda and a to-do list?

A to-do list is written by a person and changes when that person updates it. An agent agenda is produced and revised by the agent itself as it works, based on the goal and on the results of each action.

Do all AI agents have an agenda?

Not always explicitly. Simple agents decide one step at a time without writing a plan. Agents for longer tasks usually keep an explicit list of steps, because it improves focus and makes their work easier to follow and resume.

How does an agent agenda relate to task decomposition?

Task decomposition is the act of breaking a goal into smaller steps. The agenda is the result of that decomposition, kept up to date as the agent executes and learns more.

Should users be able to edit an agent's agenda?

In many business uses, yes. Letting a person review or adjust the plan before sensitive steps run is a simple and effective form of human oversight.

Where it shows up at Thinkia

Products

Further reading

Sources

Updated: 2026-10-21

Want to see how this applies to your organisation?

Talk to an AI Expert