Skip to main content

Agent Run History Growth+This feature requires the Growth+ plan or higher

Every time a Dash Agent runs — whether triggered manually, by a schedule, or by an event — EquipDash records the full details of that run. The run history lets you see what the agent did, how long it took, how many credits it used, and whether it succeeded or needs your attention.

Viewing the Run History​

  1. Go to the Agents page.
  2. Click on the agent you want to inspect.
  3. The agent's detail page opens, showing its configuration and a Run History section.
  4. The run history lists all runs for this agent, with the most recent at the top.

The agent run history page showing a list of runs with statuses and a selected run's details

Understanding Run Statuses​

Each run has a status that tells you where it got to. A run is only marked Completed if it actually finished the job — a run that stopped part-way is always labeled honestly, never given a green tick.

StatusMeaning
RunningThe agent is working right now.
Awaiting approvalThe agent paused on an action that asks you first and is waiting for someone to approve or skip it.
CompletedThe agent finished its work. You can read the summary and every step it took.
Nothing needed doing this timeThe agent checked, found nothing to act on, and stopped. This costs nothing — no AI is used.
Stopped earlyThe agent ran out of steps before finishing. The steps it did complete are listed. Raise Max steps in the builder if this keeps happening.
Stopped — got stuckThe agent started repeating itself and was stopped. Nothing further was changed.
FailedSomething went wrong. A message explains what.
CancelledSomebody stopped the run.
Timed outThe run took too long and was stopped automatically.

Status Colors​

  • Running — blue
  • Awaiting approval — amber
  • Completed — green
  • Nothing needed doing this time — grey
  • Stopped early / Stopped — got stuck — amber
  • Failed / Timed out — red
  • Cancelled — grey

Filtering Runs by Status​

If you are looking for a specific type of run (e.g., only failed runs or runs awaiting approval):

  1. On the run history page, use the status filter dropdown at the top.
  2. Select a status to filter by (e.g., Awaiting Approval, Failed, Completed).
  3. The list updates to show only runs matching the selected status.
  4. Clear the filter to see all runs again.

Viewing Run Details​

Click on any run in the list to see its full details. The detail panel shows:

Summary​

The agent generates a plain-language summary of what it accomplished during the run. For example: "Found 3 overdue rental returns. Sent reminder emails to 2 customers. 1 customer was already contacted today and was skipped."

Run Information​

FieldDescription
Triggered byHow the run was started — Manual (by a specific user), Schedule, or Event
Started atWhen the run began
Completed atWhen the run finished (if finished)
DurationHow long the run took
Credits usedThe number of AI credits consumed by this run

Actions​

The actions section lists everything the agent did, in order, written as plain sentences — "Emailed the customer about their overdue return", not a technical function name. Each step also shows whether it went ahead, is waiting for you, or was skipped, and the time it happened.

Click a step to expand it if you want the exact detail behind the sentence — what the agent asked for, and what came back.

A run written as a plain-English list of steps, with the one waiting on your OK opened up

Show Dash's thinking​

If you want to understand why the agent did something, flip the Show Dash's thinking switch above the list. The agent's reasoning appears between the steps, in its own words. Leave it off for the short version.

For actions that asked you first, you will also see who responded and when.

Telling Dash it got something wrong​

If a step is not what you wanted, you can teach the agent instead of just fixing it by hand:

  1. Expand the step.
  2. Click Not what you wanted?
  3. Type what Dash should have done instead.
  4. Click Send to Dash.

Your note feeds into the suggestions on the Dash Skills page, where you decide whether to keep it. Nothing changes until you do.

Learn more about Dash Skills

Managing Pending Approvals​

If a run is Awaiting approval, there is an action waiting for you.

The quickest place to deal with it is the Waiting for your OK box at the top of the Agents page — it collects pending actions from every agent so you can clear them in one pass. You can also respond inside the run itself.

Approving an action​

  1. Open the run with Awaiting approval status.
  2. Find the pending step.
  3. Read the sentence describing what the agent wants to do, and expand it if you want more detail.
  4. Click Approve to let it go ahead — or Approve & don't ask again to also stop being asked about that same kind of action for the rest of this run.

If several actions are waiting in the same run, Approve all clears them together.

Skipping an action​

  1. Open the run with Awaiting approval status.
  2. Find the pending step.
  3. Click Skip.

Skipping does not cancel the run. Only that one action is passed over; the agent carries on with everything else it had planned. The skip is recorded, along with who skipped it and when.

Cancelling a run​

You can stop a run at any time while it is Running or Awaiting approval.

  1. Open the run.
  2. Click Cancel run.
  3. The run stops immediately and is marked Cancelled.

Checking What a Run Cost​

Each run shows what it used:

  • Tokens used — the raw amount of AI work the run needed.
  • Share of your monthly allowance — a plain percentage, so you can see at a glance whether a run is cheap or expensive.

A run that finds nothing needed doing this time uses no AI at all.

To see usage across all your agents, go to Settings > AI Assistant.

Following Up on a Run​

After a run completes, you can send it back out with extra instructions instead of starting from scratch. On the run page, type what you want in the Follow up box and click Run — the new run starts with the previous run's results as context.

Troubleshooting​

Stopped early — the agent ran out of steps. Raise Max steps in the builder, or narrow the instructions so there is less to get through.

Stopped — got stuck — the agent kept trying the same thing. This usually means the instructions are ambiguous, or the agent was asked to do something it does not have the right switch turned on for. Check step 2 of the builder and tighten the instructions.

Waiting to start — the run is queued behind other background work. If it sits there for more than a few minutes, stop it and try again later.

Failed — open the run to see the message. Common causes:

  • Out of AI credits. Buy a credit pack from Settings > Billing & Plan.
  • Something the agent looked for no longer exists (for example, a booking that was deleted). Adjust the instructions if it keeps happening.

Timed out — the task was too broad. Narrow the instructions or reduce how much the agent has to work through.