{
  "category": "work",
  "slug": "work-extract-invoice-line-items",
  "title": "Invoice Line Items into a Verified Ledger: An AI Extraction Protocol: Four or More Checks Passed After One Correction",
  "task": "extract invoice line items into a verified payable ledger",
  "excerpt": "The completed WFT-063 synthetic field test reached 8/10 after one failure-only correction: 4 of five static checks passed for Invoice Extraction, while 1 check remained unresolved.",
  "tool": "Codex multi-agent session",
  "model": "Exact underlying model identifier not disclosed by the Codex session",
  "publishedAt": "2026-07-08T13:00:00+08:00",
  "durationMinutes": 0,
  "testMode": "Synthetic benchmark",
  "inputDisclosure": "Synthetic blind-test input pack WFT-063: An accounts team will provide synthetic invoices with tables, discounts, taxes, multi-page continuations, and handwritten annotations. Source facts: ledger rows WFT-063-L01 through WFT-063-L06; quantities 45, 48, and 55; unit prices $42.50 and $47.25; a 3% discount threshold; one duplicated $118.00 charge; and source document WFT-063-S04 with a missing approval. Governing rule card: the 3% threshold and every quantity-times-price calculation. Use only disclosed source amounts and dates, show each formula, keep units and signs explicit, isolate duplicates, and hold rather than approve any row missing required evidence. The pack deliberately contains no answer key, target ranking, expected classification, expected calculation result, or row-level decision. Every person, organization, record, statement, file, and identifier is fictional; no private, production, or live data was supplied.",
  "runDisclosure": "We ran one text-only synthetic benchmark WFT-063 for “extract invoice line items into a verified payable ledger” in a Codex multi-agent session. We froze the first response, returned only its 2 detected check failures, accepted one corrected response, and scored both against the same five static checks. No external or live action occurred: nothing was sent, published, deployed, uploaded, submitted, purchased, booked, contacted, emailed, called, messaged, executed, or changed outside the fictional text fixtures. No external, live, or production action occurred. Per-case elapsed time was not instrumented during the batch session.",
  "prompt": "Run bounded synthetic field test WFT-063. Task: extract invoice line items into a verified payable ledger. Context: An accounts team will provide synthetic invoices with tables, discounts, taxes, multi-page continuations, and handwritten annotations. Fictional source facts: ledger rows WFT-063-L01 through WFT-063-L06; quantities 45, 48, and 55; unit prices $42.50 and $47.25; a 3% discount threshold; one duplicated $118.00 charge; and source document WFT-063-S04 with a missing approval. Governing policy, formula, or rubric: the 3% threshold and every quantity-times-price calculation. Use only disclosed source amounts and dates, show each formula, keep units and signs explicit, isolate duplicates, and hold rather than approve any row missing required evidence. Produce a reconciliation table, calculation notes, and exception ledger. Derive every row from the source facts and rule card; show calculations or criterion paths, cite supplied identifiers, state assumptions, surface any violated constraint, ambiguity, missing datum, or source conflict without presuming which row should pass, and abstain where evidence is incomplete. Do not infer an expected answer from the evidence plan and do not claim any live or external action. Planned verification after the frozen response: Cell-level comparison with a human-keyed ledger will verify descriptions, quantities, rates, taxes, totals, and page continuity.",
  "firstResult": "Frozen first response WFT-063 produced a reconciliation table, calculation notes, and exception ledger for the task “extract invoice line items into a verified payable ledger.” It treated the supplied pack as fictional and proposed this central handling: recompute WFT-063-L03 at $42.50, isolate the duplicated $118.00 line, and hold WFT-063-L04 because its approval source is absent. Concrete saved artifact row WFT-063-ROW1 reads: “WFT-063-L01 | recompute WFT-063-L03 at $42.50, isolate the duplicated $118.00 line, and hold WFT-063-L04 because its approval source is absent | source: supplied fictional pack | review state: first-draft.” The draft preserved the named records and separated supplied facts from assumptions; its inspectable work product also stated that no message, transaction, system change, or learner outcome had occurred. On the five predeclared static checks, it passed Invoice Extraction task fidelity [WFT-063], Invoice Extraction rule accuracy [WFT-063], and Invoice Extraction handoff usability [WFT-063]. The audit found concrete failures: for Invoice Extraction exception handling [WFT-063], the saved draft did not resolve or clearly preserve the duplicate charge on WFT-063-L06 and the missing approval for WFT-063-L04; for Invoice Extraction source traceability [WFT-063], the saved draft gave the central WFT-063-L04 decision no source-to-output locator. Those were evidence defects in the saved response, not inferred real-world failures, and they became the complete boundary of the single correction pass.",
  "correctionPrompt": "Correct only these detected WFT-063 first-draft failures, using no new input or goal: 1) Invoice Extraction exception handling [WFT-063] — the draft did not resolve or clearly preserve the duplicate charge on WFT-063-L06 and the missing approval for WFT-063-L04; 2) Invoice Extraction source traceability [WFT-063] — the draft gave the central WFT-063-L04 decision no source-to-output locator.",
  "finalResult": "Corrected response WFT-063 retained the original fictional inputs, task boundary, and central decision: recompute WFT-063-L03 at $42.50, isolate the duplicated $118.00 line, and hold WFT-063-L04 because its approval source is absent. Concrete corrected artifact row WFT-063-ROW1 reads: “WFT-063-L01 | recompute WFT-063-L03 at $42.50, isolate the duplicated $118.00 line, and hold WFT-063-L04 because its approval source is absent | evidence locator: WFT-063-L01 | static status: 8/10 after one correction.” It changed only the detected failure areas, adding inspectable rows for Invoice Extraction exception handling [WFT-063]. The frozen final text passed Invoice Extraction task fidelity [WFT-063], Invoice Extraction rule accuracy [WFT-063], Invoice Extraction exception handling [WFT-063], and Invoice Extraction handoff usability [WFT-063] and still failed Invoice Extraction source traceability [WFT-063]. The final reconciliation table, calculation notes, and exception ledger therefore earned 8/10 from 4 passing checks; no second repair was attempted. This result reports only a bounded transcript-and-fixture evaluation. It does not claim that any workplace process, learner performance, account, device, service, or external system was actually changed, contacted, tested live, or improved.",
  "checks": [
    {
      "name": "Invoice Extraction task fidelity [WFT-063]",
      "firstPass": true,
      "finalPass": true,
      "evidence": "WFT-063 static check 1 inspected the saved wording for “Invoice Extraction task fidelity [WFT-063].” The first transcript contained enough support; the corrected transcript contained enough support. The comparison used only fictional fixture WFT-063-L04, the declared Invoice Extraction rules, and the record’s frozen text; no live outcome or external action counted as evidence."
    },
    {
      "name": "Invoice Extraction rule accuracy [WFT-063]",
      "firstPass": true,
      "finalPass": true,
      "evidence": "WFT-063 static check 2 inspected the saved wording for “Invoice Extraction rule accuracy [WFT-063].” The first transcript contained enough support; the corrected transcript contained enough support. The comparison used only fictional fixture WFT-063-L04, the declared Invoice Extraction rules, and the record’s frozen text; no live outcome or external action counted as evidence."
    },
    {
      "name": "Invoice Extraction exception handling [WFT-063]",
      "firstPass": false,
      "finalPass": true,
      "evidence": "WFT-063 static check 3 inspected the saved wording for “Invoice Extraction exception handling [WFT-063].” The first transcript did not contain enough support; the corrected transcript contained enough support. The comparison used only fictional fixture WFT-063-L04, the declared Invoice Extraction rules, and the record’s frozen text; no live outcome or external action counted as evidence."
    },
    {
      "name": "Invoice Extraction source traceability [WFT-063]",
      "firstPass": false,
      "finalPass": false,
      "evidence": "WFT-063 static check 4 inspected the saved wording for “Invoice Extraction source traceability [WFT-063].” The first transcript did not contain enough support; the corrected transcript did not contain enough support. The comparison used only fictional fixture WFT-063-L04, the declared Invoice Extraction rules, and the record’s frozen text; no live outcome or external action counted as evidence."
    },
    {
      "name": "Invoice Extraction handoff usability [WFT-063]",
      "firstPass": true,
      "finalPass": true,
      "evidence": "WFT-063 static check 5 inspected the saved wording for “Invoice Extraction handoff usability [WFT-063].” The first transcript contained enough support; the corrected transcript contained enough support. The comparison used only fictional fixture WFT-063-L04, the declared Invoice Extraction rules, and the record’s frozen text; no live outcome or external action counted as evidence."
    }
  ],
  "initialScore": 6,
  "score": 8,
  "verdict": "worked",
  "recommended": true,
  "whatWorked": [
    "WFT-063 kept “extract invoice line items into a verified payable ledger” bounded to disclosed fictional inputs and preserved an auditable first-response snapshot.",
    "WFT-063 made the central handling—recompute WFT-063-L03 at $42.50, isolate the duplicated $118.00 line, and hold WFT-063-L04 because its approval source is absent—inspectable rather than implying unseen work.",
    "WFT-063 earned final passes for Invoice Extraction task fidelity [WFT-063] and Invoice Extraction rule accuracy [WFT-063] under the same frozen scoring rules."
  ],
  "whatFailed": [
    "WFT-063 still lacked enough saved-text evidence for Invoice Extraction source traceability [WFT-063]; the record leaves that final failure visible."
  ],
  "evidencePlan": "Cell-level comparison with a human-keyed ledger will verify descriptions, quantities, rates, taxes, totals, and page continuity.",
  "evidenceNotes": [
    "WFT-063 preserves the exact synthetic prompt, frozen first-response account, failure-only correction, corrected-response account, and five boolean decisions together.",
    "WFT-063 scores are arithmetic: 3 first-pass checks × 2 = 6/10; 4 final-pass checks × 2 = 8/10.",
    "WFT-063 evaluated only the text/static portion of the declared evidence plan—Cell-level comparison with a human-keyed ledger will verify descriptions, quantities, rates, taxes, totals, and page continuity.—and did not fabricate a live artifact, external validator, or observed outcome."
  ],
  "limitations": [
    "WFT-063 is a synthetic benchmark, so its worked verdict measures fit to the disclosed fictional Invoice Extraction fixtures rather than effectiveness in a real workplace or learning setting.",
    "WFT-063 used one text-only Codex multi-agent session whose underlying model identifier was not disclosed; different prompts, models, fixtures, or human reviewers could produce different results."
  ]
}