Spending by accounting status

# Instructions — Breakdown by bookkeeping status

## ⚠️ Missing rows = zero in the period, not universal non-existence

`spendesk_analyze_spend` only displays statuses with payables in the defined period.

Example: if all payables in January 2026 are `exported`, the response returns:
```json
[{ "key": "exported", "amount": ..., "share": 100 }]
```
→ `toPrepare` and `toExport` are **missing** from the result.
→ This means: **0 payables with a payableDate in Jan 2026** have these statuses — not that the account is clean.

Payables with `toPrepare`/`toExport` may exist with a `payableDate` outside the period.

## ⚠️ spendesk_analyze_spend filters by payableDate

`spendesk_analyze_spend` always uses `payableDate` as the period. To find `toPrepare` payables without a date constraint, use `get_payables` with `bookkeepingStatus=toPrepare` and no date filter.

## The 3 bookkeeping statuses

| Status | Meaning | Action |
|--------|--------------|--------|
| `toPrepare` | Not yet reviewed — awaiting accounting review | Review manually in Spendesk |
| `toExport` | Reviewed, ready to send to the ERP | **Export** (never omit `toExport` from an ERP export) |
| `exported` | Already sent to the ERP | **Exclude** from exports to avoid double-counting |

→ For an ERP export: `bookkeepingStatus = ["toExport"]` only (see S10).

## Step 1 — Breakdown over a period

```json
{
  "companyId": "...",
  "fromDate": "YYYY-MM-DD",
  "toDate": "YYYY-MM-DD",
  "groupBy": ["bookkeepingStatus"]
}
```

→ Returns the statuses present in the period, with amount and count for each.

## Step 2 — Interpreting the results

Each row:
- `key`: `"toPrepare"` | `"toExport"` | `"exported"`
- `amount`: total of payables with this status in the period
- `count`: number of payables
- `share`: % of the period total

If a row is absent → 0 payables with that status in the period (this is NOT a global guarantee).

## Step 3 — Detecting pending payables outside the period

To make sure there is no backlog, always complement with:
```json
// get_payables — no date filter
{
  "filters": { "field": "bookkeepingStatus", "operator": "=", "value": ["toPrepare", "toExport"] }
}
```
→ Returns ALL payables not yet exported, regardless of their `payableDate`.

## Live example (H1 2026 and Jul-25→Jun-26)

```
fromDate: 2026-01-01 / toDate: 2026-06-30
→ exported : EUR 8,174,523.65 (3,937 payables, 100%)
→ toPrepare : [absent — 0 payables in this period]
→ toExport  : [absent — 0 payables in this period]
```

```
fromDate: 2025-07-01 / toDate: 2026-06-30
→ exported : EUR 16,524,574.27 (7,199 payables, 100%)
→ toPrepare : [absent — 0 payables in this period]
→ toExport  : [absent — 0 payables in this period]
```

→ **Correct interpretation**: On this account, all payables with a payableDate in these periods are already `exported`. Payables with `toExport` do exist (3 were found by S10), but their `payableDate` is after June 30, 2026.

## Useful combination: status × spend type

To analyze the accounting backlog by type:
```json
{
  "groupBy": ["bookkeepingStatus", "payableType"],
  "fromDate": "...",
  "toDate": "..."
}
```
→ Lets you see whether the `toPrepare` backlog is mostly expense claims (`expenseClaim`), invoices (`invoicePurchase`), etc.

## Example prompts

- "How many invoices are still awaiting review?" → groupBy=["bookkeepingStatus"], filter toPrepare
- "What % of spend is already exported to accounting?" → groupBy=["bookkeepingStatus"], read the share of "exported"
- "Are there any invoices ready to export?" → groupBy=["bookkeepingStatus"] or get_payables with toExport
- "Accounting workflow breakdown for Q1 2026" → groupBy=["bookkeepingStatus", "payableType"]