AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Graded 16 November 2026

Can Extend AI keep corrected data in the downstream record?

Extend AI delivered the corrected receipt number through its API. The subtotal, tax and total due stayed at $85.00, $5.10 and $90.10.

1 of 1 test case passed

Every test case in this scenario has a result.

Pass rate100%1 of 1 gradable
Coverage1 of 1test case with a result
1PassDid everything it was expected to do.
0FailNo test case failed.
0Not gradableEvery result here could be graded.
0UntestedEvery test case in this scenario has a result.
The pass rate is a summary. The evidence is the test case below: what we sent, what we checked, what the tool returned, and the proof.

The test case

Each test case is judged on its own: Pass, Fail, Not gradable, or Untested. The scenario result above is based on this test case.

Correction survives the delivery pathPassEvidence
What we sent

We ran the service receipt through an Extend workflow and replaced the receipt number printed on the document in Human Review before approving the run. We then retrieved the same record through Extend's read API.

InputDocument (PDF)
Input PDF: the service receipt before correction
Open the PDF ↗
What the tool returned
Text
Delivered output
"output":  {
"value":  {
"tax":  "$5.10",
"subtotal":  "$85.00",
"total_due":  "$90.10",
"receipt_number":  "RHS-99999"
},
"metadata":  {

}
},

Initial extraction
"initialOutput":  {
"value":  {
"tax":  "$5.10",
"subtotal":  "$85.00",
"total_due":  "$90.10",
"receipt_number":  "RHS-70234"
},

Reviewed output
"reviewedOutput":  {
"value":  {
"tax":  "$5.10",
"subtotal":  "$85.00",
"total_due":  "$90.10",
"receipt_number":  "RHS-99999"
},
"metadata":  {

}

Edit history
"edits":  {
"receipt_number":  {
"notes":  "Benchmark test: receipt_number deliberately changed from RHS-70234 to RHS-99999 to check whether the corrected value reaches the saved run record.",
"fieldType":  "string",
"editedValue":  "RHS-99999",
"originalValue":  "RHS-70234"
}
},

API response: the corrected receipt number delivered to downstream readers

Copied from Output file (JSON), lines 638–648

Open the JSON file ↗

Copied from Output file (JSON), lines 649–655

Copied from Output file (JSON), lines 707–716

Copied from Output file (JSON), lines 729–736

Expected vs. Found
✓Expected: A downstream system must receive the corrected receipt number through one of Extend's own delivery paths.Found: The API's main output and reviewed output contained the corrected receipt number; the initial extraction retained the original number.
Supporting proof
Proof 1Screenshot
Screenshot: the original receipt number before editing
Open original ↗
Proof 2Screenshot
Screenshot: the replacement receipt number entered in Human Review
Open original ↗
Proof 3Screenshot
Screenshot: original and reviewed receipt values after approval
Open original ↗
Why this result

The correction reached the output a downstream system reads, not just the review screen. That meets delivery through one of Extend's own paths; the demonstrated scope is its read API.

Also observed in this test case

Processing and read-back costs. The receipt workflow used 5 credits in total. The API call that retrieved it was not billed.

Tested by Rugved Nichite · evidence dated 16 October 2026

Configuration and setup

How this tool was set up for the run and what the test needed in place. Each row is a fact from the run's records; a fact the records do not hold is left out, not guessed.

Software that produced the output
Extend AI
Build or version
Extend API version 2026-02-09 (workflow version: draft; Parse engine 2.1.0)
Plan or tier
Pay As You Go, free level (credits remaining)
Surface
API
Set up before the run
The service receipt's receipt number had to be corrected in Human Review, with the same record then read through Extend's API in the same session.
Tested
16 October 2026 · Rugved Nichite

How this scenario is graded

How we decide Pass, Fail and Not gradable. The same rules apply to every tool tested on this scenario.

How results are decided

Each test case gets one result per tool: Pass, Fail or Not gradable. A test case we haven't run yet shows Untested. There are no partial results.

The rules
  • Pass: the tool did everything the test expected, and nothing it said contradicts the correct answer.
  • Fail: at least one expected behaviour clearly didn't happen; the test case's result says which one and quotes the tool.
  • Not gradable: our evidence couldn't settle the outcome (for example a record we needed is missing). It is never counted as a fail, and the test case's result says what's missing.

Where this sits in the benchmark

This page is one cell of a larger study: one tool, one scenario. Only this benchmark's frame appears here.

History of this result

What has happened to this result since it was first published. Runs and grades are never overwritten: a retest or a re-grade publishes a new result and keeps the earlier one readable.

from the publication record
17 November 2026First publishedStructured Document Extraction v1

Act on this result

Nothing filed here edits the run or the grade. A challenge opens a review, and a review can produce a new run or a re-grade — which becomes the current result and leaves this one in the history.

This matches what I see

You run the same kind of test against your own setup and get the same behaviour.

Agree →
This does not match

Yours behaves differently. Tell us what you got, with a screenshot if you have one.

Disagree →
Point out an issue

Something here is wrong — a reference value, a transcription, a grade.

Report an issue →
Request a retest

On a newer build, a larger dataset, or your own setup.

Request a retest →
We have fixed this

Tell us what changed and we schedule a rerun of the failing test case. The old result stays as history.

Vendor notice →
Filed against this evidenceNothing yet. Challenges, counter-evidence and fix notices appear here with their outcome, and stay on the page after they are resolved.
Cite this result
aidemos.com/benchmarks/structured-document-extraction/results/extend-ai/corrected-data-goes-downstream · 1 pass · 0 fail · coverage 1/1 · graded 2026-11-16

The same record is available as structured data through the AI Demos MCP server, with the counts, the coverage and every per-test-case reason carried as fields.

Verify the proof files

These files support this result. Open a file to inspect the original evidence.

Input · Document (PDF) ↗application/pdf · 54 KB
Output file (JSON) ↗text/plain; charset=utf-8 · 96 KB
Proof 1 · Screenshot ↗image/png · 156 KB
Proof 2 · Screenshot ↗image/png · 156 KB
Proof 3 · Screenshot ↗image/png · 181 KB
File fingerprints (SHA-256)

A fingerprint identifies the exact file used for this result.

Input · Document (PDF)a96bbe1c5b22336e7b54169baf72b85e45eabf1312cd293a1821017b52fa85f7
Output file (JSON)e7d73513c1ab29ce03e31b4d97d39a828441955690ba785fd6df16813f54e427
Proof 1 · Screenshotc1404bcc388b128ab799d43823b32a8706d5200331aca06215478ee54972fc76
Proof 2 · Screenshot2ecab40df25238de202b6ad06c3ab5e54361b1842c3acd0f6f2b7633bd096b4f
Proof 3 · Screenshotc8145e7894ca41f97863930ca8711a99ddd879aa3365101ce0bacaf1a8358726
Same content as Output file22072ba4c12f4a9b78ee00bc3cc20ddacd3cba1571a9326ff4d430b088814276