> For the complete documentation index, see [llms.txt](https://docs.neoflo.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.neoflo.ai/invoice-processing/memory-and-reporting.md).

# Memory & Reporting

Manage the platform's learned field corrections, track volume, SLAs, and stage health, and report issues to the team.

### Memory <a href="#id-98a8ab47-b9a6-473f-8453-4035bda0f596" id="id-98a8ab47-b9a6-473f-8453-4035bda0f596"></a>

Memory is how the platform learns from your corrections. Every time you fix a field, the platform records the correction. Once it sees the same correction enough times, it starts applying it automatically.

The Memory page lists **Field Corrections**.

<table data-header-hidden><thead><tr><th width="167.57421875"></th><th></th></tr></thead><tbody><tr><td><strong>Column</strong></td><td><strong>What it shows</strong></td></tr><tr><td><strong>Field</strong></td><td>The field the correction applies to (for example, <code>vendor_name</code> or <code>customer_legal_entity</code>).</td></tr><tr><td><strong>Details</strong></td><td>The mapping the platform learned, from the raw extracted value to the corrected value.</td></tr><tr><td><strong>Streak</strong></td><td>How many consecutive times the correction has been confirmed (for example, 1x, 2x, 3x).</td></tr><tr><td><strong>Status</strong></td><td><strong>Learning</strong> while the streak builds, or <strong>Auto-applying</strong> once the platform applies it on its own.</td></tr><tr><td><strong>Actions</strong></td><td>Delete an entry to restore manual review for that field.</td></tr></tbody></table>

A correction stays in **Learning** while its streak builds. Once it reaches the auto-apply threshold, its status becomes **Auto-applying** and the platform corrects that field for you going forward. Delete the entry any time to send the field back to manual review.

### Reporting <a href="#db4cbe07-0f0c-4b94-a9fe-ce18545f5daa" id="db4cbe07-0f0c-4b94-a9fe-ce18545f5daa"></a>

The Reporting workspace answers four questions: are we processing the volume we expected, where are invoices stuck, is the extraction accurate, and are we faster than the manual baseline.

Two controls at the top filter everything on the page:

* **Range**: Today, 7d, 30d, or a Custom range.
* **Entity**: All, or a specific entity: SG, MY, HK, ID, PH, or other.

#### 1. Volume & Coverage <a href="#id-16e0b377-1609-43ba-bc3b-499580409388" id="id-16e0b377-1609-43ba-bc3b-499580409388"></a>

The headline cards track throughput against target.

<table data-header-hidden><thead><tr><th width="181.12890625"></th><th></th></tr></thead><tbody><tr><td><strong>Metric</strong></td><td><strong>What it measures</strong></td></tr><tr><td><strong>Invoices posted</strong></td><td>Successfully posted to the ERP in the selected window.</td></tr><tr><td><strong>Invoices created</strong></td><td>Top-of-funnel demand, invoices that entered the platform.</td></tr><tr><td><strong>Coverage</strong></td><td>Posted as a share of the monthly target.</td></tr><tr><td><strong>Avg SLA (p50)</strong></td><td>Median time from Review clicked to ERP posting.</td></tr></tbody></table>

Below the cards: a daily trend of created vs. posted volume against a per-day target line, a **Posted by entity** breakdown, a **Funnel** showing the count at each stage (Extraction, Matching, ERP Posting, Posted), and a **Drop-off by reason** panel showing where invoices exit the funnel.

#### 2. Stage Health <a href="#id-04ad9b9d-fff2-4d62-bfb1-9ac04fbb65c3" id="id-04ad9b9d-fff2-4d62-bfb1-9ac04fbb65c3"></a>

A live snapshot of where invoices are currently stuck, broken out by entity. Each panel flags invoices sitting longer than expected.

<table data-header-hidden><thead><tr><th width="205.1328125"></th><th></th></tr></thead><tbody><tr><td><strong>Panel</strong></td><td><strong>Flag threshold</strong></td></tr><tr><td><strong>Stuck in extraction</strong></td><td>More than 5 minutes in the extraction stage.</td></tr><tr><td><strong>Stuck in matching</strong></td><td>More than 15 minutes on the summary screen.</td></tr><tr><td><strong>Stuck in ERP posting</strong></td><td>More than 2 minutes in ERP posting after matching is confirmed.</td></tr></tbody></table>

#### 3. Quality & Accuracy <a href="#id-0d4c52d8-e5b6-4b72-85f6-5fa8c95fd9b4" id="id-0d4c52d8-e5b6-4b72-85f6-5fa8c95fd9b4"></a>

The **Rejection split** shows user-initiated rejections broken down by reason, so you can see whether the extraction is doing its job and where corrections cluster.

//need quality & accuracy screenshot with data

#### 4. Efficiency & SLA <a href="#id-785c6d8b-ae0c-4c4c-97ff-21d6bc017e7a" id="id-785c6d8b-ae0c-4c4c-97ff-21d6bc017e7a"></a>

This section compares processing speed against the manual baseline of roughly 6 minutes per invoice.

<table data-header-hidden><thead><tr><th width="249.40625"></th><th></th></tr></thead><tbody><tr><td><strong>Metric</strong></td><td><strong>What it measures</strong></td></tr><tr><td><strong>End-to-end SLA, percentiles</strong></td><td>Total time from when the invoice entered the platform to when it posted, including any time it sat unassigned or unopened.</td></tr><tr><td><strong>Review-to-posted time, percentiles</strong></td><td>Only the time from Review clicked to posted. Strips out queue and wait time, this is how fast a human moves once they've started.</td></tr><tr><td><strong>SLA trend (p50 vs. p95)</strong></td><td>The typical case (p50) tracked against the worst case (p95) over time, so you can see if performance is drifting or holding steady.</td></tr><tr><td><strong>Per-agent throughput</strong></td><td>Invoices posted per agent, colored by median review-to-posted time.</td></tr><tr><td><strong>Auto-confirm rate by stage</strong></td><td>Of all invoices, the share that passed automatically at each individual stage, extraction, matching, or posting, without a manual correction.</td></tr><tr><td><strong>Full auto-confirm (STP-eligible)</strong></td><td>The share of invoices that passed automatically at all three stages combined, the true straight-through-processing rate. See <a href="https://app.gitbook.com/o/GpIuVfNw48gFPE2CRQKe/sites/site_KxN62/s/gsLN5wIf4V78YgAzUfqh/~/edit/~/changes/4/invoice-processing/processing-an-invoice#id-82ff60db-6bf7-4c94-9724-7fc483ce39ea">Straight-Through Processing</a> for what that means.</td></tr></tbody></table>

### Reporting An Issue <a href="#id-65ac8488-5a48-4980-ad85-6a0012ff1e6d" id="id-65ac8488-5a48-4980-ad85-6a0012ff1e6d"></a>

If you run into a problem, click **Report Issue** in the left navigation.

{% stepper %}
{% step %}

#### Select your issue type

Choose an issue type from the dropdown.
{% endstep %}

{% step %}

#### Describe what happened

Enter the details in the text box.
{% endstep %}

{% step %}

#### Attach a screenshot (optional)

Add a screenshot or short clip: JPG, JPEG, PNG, BMP, WEBP, or MP4, up to 1MB each, 5MB total.
{% endstep %}

{% step %}

#### Submit

Click **Submit** to send the report to the team. You'll get an email confirmation with a ticket ID for tracking.
{% endstep %}
{% endstepper %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.neoflo.ai/invoice-processing/memory-and-reporting.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.