> For the complete documentation index, see [llms.txt](https://inistate.gitbook.io/home/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://inistate.gitbook.io/home/mcp-integration/working-with-entries.md).

# Working with Entries

Entry operations are available in all three modes. The AI uses the `execute_activity` and `diagnose_entry` guided prompts to walk through the correct tool sequence automatically.

## List and Filter Entries

Ask the AI to query entries in plain language. You can filter by state, field values, or free text.

{% tabs %}
{% tab title="Example prompts" %}

> "Show me all open purchase orders."

> "List invoices created this week sorted by amount."

> "Find all entries assigned to me in the Leave Request module."
> {% endtab %}

{% tab title="What the AI does" %}

1. Calls `list_modules` to confirm the module name
2. Calls `list_entries` with your filters, sort, and pagination settings
3. Returns a summarized list with key fields

Filters support equality, ranges, contains, date comparisons, and more. The AI translates your natural language into the correct filter syntax.
{% endtab %}
{% endtabs %}

## View a Single Entry

{% tabs %}
{% tab title="Example prompts" %}

> "Show me entry INV-1042."

> "What's the current state of leave request LR-0088?"
> {% endtab %}

{% tab title="What the AI does" %}

1. Calls `get_entry` with the module name and entry ID
2. Returns current field values, state, and available next activities
   {% endtab %}
   {% endtabs %}

## Create an Entry

{% tabs %}
{% tab title="Example prompts" %}

> "Create a new purchase order for supplier Acme Corp, amount RM 5,000."

> "Raise a leave request for John Tan from 10 June to 14 June."
> {% endtab %}

{% tab title="What the AI does" %}

1. Calls `get_form` for the `create` activity to discover required fields and their types
2. Maps your details to the form fields
3. Calls `submit_activity` with the prepared input and an AI traceability record
4. Confirms the new entry ID
   {% endtab %}
   {% endtabs %}

{% hint style="warning" %}
The AI will always call `get_form` before submitting. If required fields cannot be confidently filled from your prompt, it will ask you for the missing information rather than guessing.
{% endhint %}

## Edit an Entry

{% tabs %}
{% tab title="Example prompts" %}

> "Update the amount on PO-0234 to RM 7,500."

> "Change the supplier on invoice INV-1042 to Beta Supplies."
> {% endtab %}

{% tab title="What the AI does" %}

1. Calls `get_form` for the `edit` activity on the specific entry
2. Applies your changes on top of the current values
3. Calls `submit_activity` to save
   {% endtab %}
   {% endtabs %}

## Delete an Entry

{% tabs %}
{% tab title="Example prompts" %}

> "Delete leave request LR-0099."
> {% endtab %}

{% tab title="What the AI does" %}
Calls `submit_activity` with activity `delete` and the entry ID. This is a permanent action.
{% endtab %}
{% endtabs %}

{% hint style="danger" %}
Deletion cannot be undone. The AI will confirm with you before proceeding.
{% endhint %}

## Run a Custom Activity

Workflow modules have custom activities such as "Approve", "Reject", "Escalate", or "Submit for Review". Ask the AI to run any of them.

{% tabs %}
{% tab title="Example prompts" %}

> "Approve leave request LR-0055."

> "Escalate PO-0234 to the finance manager."
> {% endtab %}

{% tab title="What the AI does" %}

1. Calls `get_form` for the named activity to discover its fields and confidence threshold
2. Prepares input and assesses its confidence
3. Calls `submit_activity`. If confidence is below the activity's threshold, the entry is flagged for human review instead of transitioning state
   {% endtab %}
   {% endtabs %}

## Bulk Operations

To create or update many entries at once, ask the AI to do them in a single batch rather than one by one.

{% tabs %}
{% tab title="Example prompts" %}

> "Create these 10 purchase orders all at once." *(with a list or table)*

> "Mark all these leave requests as approved."
> {% endtab %}

{% tab title="What the AI does" %}
Calls `submit_activities` with all items in a single request instead of making a separate `submit_activity` call for each. Returns per-item results so you can see which succeeded and which failed.

Max 100 items per batch. For larger sets the AI will chunk automatically.
{% endtab %}
{% endtabs %}

## View Entry History

Every entry has a full audit trail of who did what and when, including field-level change details and the AI reasoning behind automated actions.

{% tabs %}
{% tab title="Example prompts" %}

> "Show me the history of invoice INV-1042."

> "Who approved leave request LR-0055 and when?"

> "Why was PO-0234 escalated?"
> {% endtab %}

{% tab title="What the AI does" %}

1. Calls `get_entry` for current state
2. Calls `get_entry_history` for the full chronological audit trail
3. Summarizes the timeline, including AI confidence scores and reasoning for automated steps
   {% endtab %}
   {% endtabs %}

## AI Traceability

Every write operation (create, edit, delete, custom activity) is automatically logged with:

* **Reasoning** — why the AI chose this action
* **Model** — which AI model performed the action
* **Confidence** — the AI's self-assessed confidence (0–1)
* **Sources** — any documents or context the AI referenced

This appears in the entry history and is visible to all workspace users.