{
  "slug": "monitoring-document-extraction-accuracy-in-production",
  "title": "Monitoring Document Extraction Accuracy in Production: Audit Samples, Field Metrics and Drift",
  "description": "How to monitor document extraction accuracy in production: a stratified audit sample with independent labels, per-field metrics with intervals, extraction-specific drift triggers, thresholds from your own variance, and regression intake.",
  "kind": "sub",
  "order": 10,
  "target_query": "monitoring document extraction accuracy in production",
  "secondary_queries": [],
  "tags": [
    "document-extraction",
    "monitoring",
    "accuracy",
    "drift",
    "production"
  ],
  "published": "2026-10-28",
  "updated": "2026-10-28",
  "words": 1449,
  "estimated_tokens": 1927,
  "premium": false,
  "rights": {
    "access": "free",
    "note": "Editorial guides are always free and never part of the licensed corpus.",
    "license": "https://changegamer.ai/license.xml",
    "pricing": "https://changegamer.ai/api/pricing.json",
    "payment": "https://changegamer.ai/api/payment.json"
  },
  "license": "https://changegamer.ai/license.xml",
  "citation": "ChangeGamer (2026-10-28). Monitoring Document Extraction Accuracy in Production: Audit Samples, Field Metrics and Drift. ChangeGamer. https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production (updated 2026-10-28).",
  "bibtex": "@misc{changegamer_monitoring_document_extraction_accuracy_in_production, title = {Monitoring Document Extraction Accuracy in Production: Audit Samples, Field Metrics and Drift}, publisher = {ChangeGamer}, year = {2026}, url = {https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production}, note = {Updated 2026-10-28}}",
  "canonical": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production",
  "markdown": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production.md",
  "takeaways": [
    "Monitoring document extraction accuracy in production means labelling a small, stratified sample of live documents on a schedule, because schema validity and pass-through checks cannot show whether a value is actually correct.",
    "Document extraction accuracy should be scored per field, not per document, and each miss classified as missing, wrong or hallucinated, since the three have different fixes (author inference).",
    "On a small audit sample, report a confidence interval beside every field accuracy, because a point estimate from a few dozen documents can hide a wide range of plausible true rates (general statistical practice).",
    "Extraction drift usually arrives through a named change, such as a vendor template revision, a parser or model pin change, or a shift in document-family mix, so each of these should trigger a fresh audit sample (author inference).",
    "Alert thresholds for extraction accuracy should be set from your own baseline labels plus a margin derived from their sample variance, because no universal accuracy figure transfers between document sets."
  ],
  "outline": [
    {
      "depth": 2,
      "text": "How do you build an audit sample from production documents?",
      "anchor": "how-do-you-build-an-audit-sample-from-production-documents",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#how-do-you-build-an-audit-sample-from-production-documents"
    },
    {
      "depth": 3,
      "text": "Why not just use the human review queue as the sample?",
      "anchor": "why-not-just-use-the-human-review-queue-as-the-sample",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#why-not-just-use-the-human-review-queue-as-the-sample"
    },
    {
      "depth": 2,
      "text": "Which field-level metrics should you track?",
      "anchor": "which-field-level-metrics-should-you-track",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#which-field-level-metrics-should-you-track"
    },
    {
      "depth": 3,
      "text": "How do you handle small samples?",
      "anchor": "how-do-you-handle-small-samples",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#how-do-you-handle-small-samples"
    },
    {
      "depth": 2,
      "text": "What does drift look like in a document pipeline?",
      "anchor": "what-does-drift-look-like-in-a-document-pipeline",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#what-does-drift-look-like-in-a-document-pipeline"
    },
    {
      "depth": 2,
      "text": "How do you set alert thresholds without a universal accuracy number?",
      "anchor": "how-do-you-set-alert-thresholds-without-a-universal-accuracy-number",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#how-do-you-set-alert-thresholds-without-a-universal-accuracy-number"
    },
    {
      "depth": 2,
      "text": "What should flow from production into the regression set?",
      "anchor": "what-should-flow-from-production-into-the-regression-set",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#what-should-flow-from-production-into-the-regression-set"
    },
    {
      "depth": 2,
      "text": "Does ChangeGamer have first-hand evidence on monitoring extraction?",
      "anchor": "does-changegamer-have-first-hand-evidence-on-monitoring-extraction",
      "url": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production#does-changegamer-have-first-hand-evidence-on-monitoring-extraction"
    }
  ],
  "faq": [
    {
      "question": "How do you measure document extraction accuracy in production without ground truth?",
      "answer": "Production document extraction accuracy is measured by creating ground truth by sampling live documents and having a person label them without seeing the model output. Stratify the sample by document family, vendor template and risk class, and add a random slice to catch strata you did not anticipate. This is general sampling practice and author design, not a rule from the reference corpus."
    },
    {
      "question": "How big should a document extraction audit sample be?",
      "answer": "A document extraction audit sample has no universal size; derive it from the precision you need per field. Compute an interval on a pilot sample and keep sampling until the interval is narrow enough to tell a real drop from noise. Rare fields and small strata need their own quota (general statistical practice, author inference)."
    },
    {
      "question": "What causes document extraction accuracy to drop after launch?",
      "answer": "Accuracy in a document extraction pipeline typically drops because of a vendor changing its template, a parser or model version change, and a shift in which document families arrive. The extraction reference treats parsing and field extraction as separate stages with separate failure modes, so check both stages (author inference on the causes)."
    },
    {
      "question": "Which production documents should become regression tests?",
      "answer": "Add documents whose audit label disagreed with the extraction, plus one representative per new template or family. Store the page image reference, the corrected record, the parser, model and schema versions, and a content hash so that duplicates are not added twice (author design)."
    }
  ],
  "body": "Monitoring document extraction accuracy in production means periodically labelling a sample of live documents by hand and scoring each field against that label. Nothing else measures field accuracy against ground truth. The [Structured Outputs and JSON Mode](/resources/structured-outputs-and-json-mode) reference (updated 2026-09-26) describes what constrained decoding gives you, a schema-valid object, and a valid object can still carry a wrong amount. The [Document Extraction and Parsing for Agents](/resources/document-extraction-for-agents) reference (updated 2026-09-28) separates parsing from field extraction because each has its own failure modes. The [AI document extraction pipeline](/articles/ai-document-extraction-pipeline-in-production) pillar covers the whole pipeline; this piece covers how to know it is still working. Anything not cited to a reference is author inference or general statistical practice, labelled as such.\n\n## How do you build an audit sample from production documents?\n\nBuild the audit sample by drawing documents from live traffic in strata, then adding a random slice, and have someone label them without seeing the extractor output. The [Evaluating AI Agents](/resources/evaluating-ai-agents) reference (updated 2026-09-27) says to evaluate your own task distribution because public benchmark scores diverge from production performance; the audit sample is that principle applied to documents. The scheme below is author design.\n\n| Slice | How to draw it | What it catches |\n|---|---|---|\n| Document family | A quota per family (invoice, contract, form) | A family that quietly degrades while the average holds |\n| Vendor or template | A quota per frequent sender or layout | A template revision at one counterparty |\n| Risk class | Oversample documents whose fields touch payments, identity or health | Errors with the highest consequence |\n| Random slice | A plain random draw across everything | Strata you did not think to define |\n\nLabel independently. The labeller should see the page image and the schema, not the extracted record, because a reviewer shown the model's answer can drift toward accepting it (author inference). Where two people label a subset, disagreement shows a field definition that is unclear, which is a schema problem, not a model problem.\n\n### Why not just use the human review queue as the sample?\n\nThe review queue is a biased sample, because documents reach it only after a trigger fires. Its accuracy says nothing about documents that passed unreviewed, and those are the ones monitoring exists to check. How triggers route and how the queue is run is in [human review fallback for document extraction](/articles/human-review-fallback-for-document-extraction); here, use the queue as a source of known failures, never as the accuracy estimate (author inference).\n\n## Which field-level metrics should you track?\n\nTrack accuracy per field, not per document, with exact and normalised match reported separately and every miss classified. A document-level score hides which field broke: one wrong date in twenty fields fails the whole document and tells you nothing about the nineteen right ones (author inference).\n\n| Metric | Definition | Why separate |\n|---|---|---|\n| Exact match | Extracted string equals the label character for character | Strict; shows formatting drift |\n| Normalised match | Equal after a stated normalisation (trimmed, dates to ISO, currency stripped) | Shows value correctness independent of format |\n| Missing | Label has a value, extraction is null or absent | Points at parsing or schema description |\n| Wrong | Both have values and they differ | Points at reading or reasoning error |\n| Hallucinated | Extraction has a value, the label says the field is absent | Points at prompt, enum and null handling |\n\nWrite the normalisation rules down and version them, otherwise a rule change looks like an accuracy change. For tables, score rows and cells the way [table and form validation for document extraction](/articles/table-and-form-validation-for-document-extraction) defines them, and note that a multimodal reader adds its own error types, covered in [multimodal document understanding for extraction pipelines](/articles/multimodal-document-understanding-for-extraction-pipelines).\n\n### How do you handle small samples?\n\nReport a confidence interval with every field accuracy, because few labelled documents leave wide uncertainty. An illustrative case: 48 correct of 50 labelled values is 96 percent as a point estimate, but a 95 percent Wilson interval runs roughly 86.5 to 98.9 percent. That range is wide enough that a fall to 92 percent next week may be noise. Wilson intervals are standard general practice and the numbers here are an arithmetic illustration, not a measurement of any pipeline. Report each stratum separately; a pooled interval can hide a weak template.\n\n## What does drift look like in a document pipeline?\n\nDocument pipelines usually drift through a named, checkable change, so list those changes and make each one trigger a fresh audit sample. Generic agent drift, meaning shifting prompts, tools and user behaviour, is covered in [online agent metrics and drift monitoring](/articles/online-agent-metrics-and-drift-monitoring). The triggers below are specific to extraction (author inference).\n\n1. **Vendor or template change.** A counterparty redesigns its invoice. Track the share of documents per template fingerprint (for example a layout hash or sender) and sample any template whose share jumps or that appears for the first time.\n2. **Parser or model pin change.** A new parser release or model snapshot changes output without any change on your side. Record both versions on every extraction and re-sample after every bump, as the version fields in the [extraction schema versioning](/articles/extraction-schema-versioning-for-document-pipelines) approach make possible.\n3. **Shift in document-family mix.** Overall accuracy can fall with no per-family change, simply because a harder family became a larger share. Monitor the mix next to the accuracy, and report per-family results.\n4. **Input quality shift.** A new scan source or capture app changes resolution or skew. Track simple input stats such as page count and image size alongside results.\n\nCompare each triggered sample against a fixed baseline window, using the baseline discipline in [online agent metrics and drift monitoring](/articles/online-agent-metrics-and-drift-monitoring); re-baseline only after a known, accepted change.\n\n## How do you set alert thresholds without a universal accuracy number?\n\nSet each threshold from your own baseline labels, because no source in the corpus publishes an acceptable extraction accuracy and a figure borrowed from another document set does not transfer (author inference). The calibration steps (labelled sample, per-field sort, stricter lines for money fields) are in the [pillar's threshold section](/articles/ai-document-extraction-pipeline-in-production), and re-calibration after a change is covered in [human review fallback for document extraction](/articles/human-review-fallback-for-document-extraction). This section adds only the alert rule for monitoring.\n\nUse the lower bound of the baseline interval as the alert line, and compare a new sample's interval against it. That rule has a power limit worth stating. With 50 labels per field the baseline lower bound is about 86.5 percent (the 48-of-50 case above), and a new sample's upper bound drops below it only at 38 of 50 correct, or 76 percent. A fall from 96 to 90 percent passes silently (arithmetic illustration, general practice). Two ways to trade this off, both author inference:\n\n- **Label more:** the interval narrows with sample size, so smaller drops become detectable.\n- **Two tiers:** page only on the strict rule, and open an investigation when the point estimate crosses the line, accepting some false alarms for earlier warning.\n\nRisk shapes the quota too: a field that moves money earns a larger audit sample than a free-text note, so its interval is narrower.\n\n## What should flow from production into the regression set?\n\nSend disagreements and new templates into the regression set, and record enough to replay them. Regression-set mechanics, meaning how to run and gate it, are in [incident-to-eval-fixture-loop](/articles/incident-to-eval-fixture-loop); this section covers intake only (author design).\n\n- **New tags only:** the stratum and the miss type (missing, wrong, hallucinated), because those tie an item back to the audit metrics above.\n- **Record shape, hashing and schema-version tagging:** follow the review-item record in [human review fallback for document extraction](/articles/human-review-fallback-for-document-extraction) and the correction migration rules in [extraction schema versioning](/articles/extraction-schema-versioning-for-document-pipelines) instead of a second format.\n- **Dedup and caps:** skip repeats by content hash, and cap how many items one template contributes so a single noisy sender does not dominate the set.\n- **Privacy:** keep only what your retention rules allow, as the pillar's launch checklist says.\n\nThe cost of labelling is a real line item; see [cost per page for document extraction](/articles/cost-per-page-for-document-extraction) for where it belongs in the worksheet.\n\n## Does ChangeGamer have first-hand evidence on monitoring extraction?\n\nAs of October 2026 ChangeGamer runs no document extraction, so it has no accuracy data to show. The one related practice is build-time contract validation: the loader in `src/lib/articles.ts` throws at build when an article has a missing or duplicate `targetQuery`. That is a structural check that catches a class of mistakes cheaply, and it says nothing about whether any article's content is right, which is the same gap between schema validity and accuracy described above.",
  "cluster": {
    "id": "document-pipelines",
    "title": "Document extraction pipelines in production",
    "description": "How to run AI document extraction as a production pipeline — OCR vs vision-model vs hybrid parsing, schema-constrained field extraction, table and form validation, human-review fallback, schema versioning, cost per page, open-source vs managed tooling and feeding a system of record — not chunking and retrieval (chunking-documents-for-rag, rag-ingestion-pipeline), not generic structured-output and tool-call mechanics (structured-outputs-vs-tool-calling, tool-calling-contracts-for-ai-agents) and not injection defense in depth (prompt-injection-defense-in-depth-for-agents).",
    "status": "complete",
    "pillar": {
      "slug": "ai-document-extraction-pipeline-in-production",
      "title": "AI Document Extraction in Production: Parse, Extract, Validate, Review",
      "description": "How to run AI document extraction as a production pipeline: separate parsing from schema-constrained field extraction, choose OCR, a vision model or a hybrid, validate values, route doubtful ones to a human, version the schema, and work out cost per page.",
      "kind": "pillar",
      "order": 0,
      "html": "https://changegamer.ai/articles/ai-document-extraction-pipeline-in-production",
      "markdown": "https://changegamer.ai/articles/ai-document-extraction-pipeline-in-production.md",
      "json": "https://changegamer.ai/api/articles/ai-document-extraction-pipeline-in-production.json"
    },
    "articles": [
      {
        "slug": "ocr-vs-vision-model-vs-hybrid-document-parsing",
        "title": "OCR vs Vision Language Model for Document Parsing: How to Choose Stage A",
        "description": "OCR vs vision language model for document parsing: a decision order keyed on residency, layout variance, volume and known schema, how a hybrid arbitrates disagreements, a per-field bake-off, failure-mode mapping, page batching and privacy.",
        "kind": "sub",
        "order": 1,
        "html": "https://changegamer.ai/articles/ocr-vs-vision-model-vs-hybrid-document-parsing",
        "markdown": "https://changegamer.ai/articles/ocr-vs-vision-model-vs-hybrid-document-parsing.md",
        "json": "https://changegamer.ai/api/articles/ocr-vs-vision-model-vs-hybrid-document-parsing.json"
      },
      {
        "slug": "schema-constrained-field-extraction-from-documents",
        "title": "Schema-Constrained Field Extraction From Documents: Settings and Failure Branches",
        "description": "Schema constrained field extraction from documents: nullable fields, enums with an escape value, evidence fields, a branch table for refusal, truncation, parse failure and empty output, a one-retry policy, and field confidence as inference.",
        "kind": "sub",
        "order": 2,
        "html": "https://changegamer.ai/articles/schema-constrained-field-extraction-from-documents",
        "markdown": "https://changegamer.ai/articles/schema-constrained-field-extraction-from-documents.md",
        "json": "https://changegamer.ai/api/articles/schema-constrained-field-extraction-from-documents.json"
      },
      {
        "slug": "table-and-form-validation-for-document-extraction",
        "title": "Table and Form Validation for Document Extraction: Row Counts, Totals and Required Fields",
        "description": "Validating table and form extraction from documents: rows split across pages, subtotal rows, exact money reconciliation, column sums, required-field maps per form type, cross-field checks and a failed-check-to-route table.",
        "kind": "sub",
        "order": 3,
        "html": "https://changegamer.ai/articles/table-and-form-validation-for-document-extraction",
        "markdown": "https://changegamer.ai/articles/table-and-form-validation-for-document-extraction.md",
        "json": "https://changegamer.ai/api/articles/table-and-form-validation-for-document-extraction.json"
      },
      {
        "slug": "human-review-fallback-for-document-extraction",
        "title": "Human Review Fallback for Document Extraction: Triggers, Capacity and the Correction Loop",
        "description": "Designing the human review fallback for document extraction: a routing-trigger table, cross-checking high-stakes fields, review capacity and backlog policy, the review item record and a correction loop that feeds regression sets.",
        "kind": "sub",
        "order": 4,
        "html": "https://changegamer.ai/articles/human-review-fallback-for-document-extraction",
        "markdown": "https://changegamer.ai/articles/human-review-fallback-for-document-extraction.md",
        "json": "https://changegamer.ai/api/articles/human-review-fallback-for-document-extraction.json"
      },
      {
        "slug": "extraction-schema-versioning-for-document-pipelines",
        "title": "Versioning a Document Extraction Schema: Migrations, Manifests and Rollback",
        "description": "How to version a document extraction schema: which edits break stored records, absent versus null fields, lazy upcast versus backfill, reprocessing from stored parse output, a version manifest, dual-read cutover, correction migration and rollback.",
        "kind": "sub",
        "order": 5,
        "html": "https://changegamer.ai/articles/extraction-schema-versioning-for-document-pipelines",
        "markdown": "https://changegamer.ai/articles/extraction-schema-versioning-for-document-pipelines.md",
        "json": "https://changegamer.ai/api/articles/extraction-schema-versioning-for-document-pipelines.json"
      },
      {
        "slug": "cost-per-page-for-document-extraction",
        "title": "Document Extraction Cost per Page: Tokens, Retries, Review and Budget Alarms",
        "description": "How to model document extraction cost per page beyond the parse fee: page-count mix versus per-page pricing, image tokens versus per-call overhead, retries by failure class, review share, batch versus sync, and a budget kill switch on cost per accepted document.",
        "kind": "sub",
        "order": 6,
        "html": "https://changegamer.ai/articles/cost-per-page-for-document-extraction",
        "markdown": "https://changegamer.ai/articles/cost-per-page-for-document-extraction.md",
        "json": "https://changegamer.ai/api/articles/cost-per-page-for-document-extraction.json"
      },
      {
        "slug": "open-source-vs-managed-document-parsing",
        "title": "Open Source vs Managed Document Parsing: A Decision Order, Operating Burden and Exit Plan",
        "description": "How to decide between self-hosted open-source parsers and a managed parsing service: residency and licence first, volume and operations capacity second, reversibility third, plus the cost of operating models, lock-in, a single-interface fallback design and a trial protocol on your own documents.",
        "kind": "sub",
        "order": 7,
        "html": "https://changegamer.ai/articles/open-source-vs-managed-document-parsing",
        "markdown": "https://changegamer.ai/articles/open-source-vs-managed-document-parsing.md",
        "json": "https://changegamer.ai/api/articles/open-source-vs-managed-document-parsing.json"
      },
      {
        "slug": "feeding-extraction-output-into-a-system-of-record",
        "title": "Writing Extracted Document Data to a System of Record: Matching, Keys, States and Replay",
        "description": "How to write extracted document data to a system of record: matching an extracted record to an existing one, retry keys versus business keys, a staged-to-committed state machine, per-record versus per-document atomicity, dead-letter queues, replay from the saved parse artifact and the write-boundary log.",
        "kind": "sub",
        "order": 8,
        "html": "https://changegamer.ai/articles/feeding-extraction-output-into-a-system-of-record",
        "markdown": "https://changegamer.ai/articles/feeding-extraction-output-into-a-system-of-record.md",
        "json": "https://changegamer.ai/api/articles/feeding-extraction-output-into-a-system-of-record.json"
      },
      {
        "slug": "multimodal-document-understanding-for-extraction-pipelines",
        "title": "Multimodal Document Understanding for Extraction: Charts, Diagrams, Photos, Stamps and Video Frames",
        "description": "How to handle non-text visuals in a document extraction pipeline: when OCR is not enough, extracting a value versus a description versus a class, region-level routing, figure-aware page packing, poor scans, video frames, image-borne injection and per-visual evaluation.",
        "kind": "sub",
        "order": 9,
        "html": "https://changegamer.ai/articles/multimodal-document-understanding-for-extraction-pipelines",
        "markdown": "https://changegamer.ai/articles/multimodal-document-understanding-for-extraction-pipelines.md",
        "json": "https://changegamer.ai/api/articles/multimodal-document-understanding-for-extraction-pipelines.json"
      },
      {
        "slug": "monitoring-document-extraction-accuracy-in-production",
        "title": "Monitoring Document Extraction Accuracy in Production: Audit Samples, Field Metrics and Drift",
        "description": "How to monitor document extraction accuracy in production: a stratified audit sample with independent labels, per-field metrics with intervals, extraction-specific drift triggers, thresholds from your own variance, and regression intake.",
        "kind": "sub",
        "order": 10,
        "html": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production",
        "markdown": "https://changegamer.ai/articles/monitoring-document-extraction-accuracy-in-production.md",
        "json": "https://changegamer.ai/api/articles/monitoring-document-extraction-accuracy-in-production.json"
      },
      {
        "slug": "splitting-long-documents-for-document-extraction",
        "title": "How to Split Long Documents for Extraction: Batches, Overlap and Reassembly",
        "description": "How to split long documents for extraction: size batches from measured tokens per page, choose section or page windows, overlap pages, carry table headers, reconcile conflicts across batches and reassemble with provenance.",
        "kind": "sub",
        "order": 11,
        "html": "https://changegamer.ai/articles/splitting-long-documents-for-document-extraction",
        "markdown": "https://changegamer.ai/articles/splitting-long-documents-for-document-extraction.md",
        "json": "https://changegamer.ai/api/articles/splitting-long-documents-for-document-extraction.json"
      }
    ]
  },
  "navigation": {
    "pillar": {
      "slug": "ai-document-extraction-pipeline-in-production",
      "title": "AI Document Extraction in Production: Parse, Extract, Validate, Review",
      "description": "How to run AI document extraction as a production pipeline: separate parsing from schema-constrained field extraction, choose OCR, a vision model or a hybrid, validate values, route doubtful ones to a human, version the schema, and work out cost per page.",
      "kind": "pillar",
      "order": 0,
      "html": "https://changegamer.ai/articles/ai-document-extraction-pipeline-in-production",
      "markdown": "https://changegamer.ai/articles/ai-document-extraction-pipeline-in-production.md",
      "json": "https://changegamer.ai/api/articles/ai-document-extraction-pipeline-in-production.json"
    },
    "previous": {
      "slug": "multimodal-document-understanding-for-extraction-pipelines",
      "title": "Multimodal Document Understanding for Extraction: Charts, Diagrams, Photos, Stamps and Video Frames",
      "description": "How to handle non-text visuals in a document extraction pipeline: when OCR is not enough, extracting a value versus a description versus a class, region-level routing, figure-aware page packing, poor scans, video frames, image-borne injection and per-visual evaluation.",
      "kind": "sub",
      "order": 9,
      "html": "https://changegamer.ai/articles/multimodal-document-understanding-for-extraction-pipelines",
      "markdown": "https://changegamer.ai/articles/multimodal-document-understanding-for-extraction-pipelines.md",
      "json": "https://changegamer.ai/api/articles/multimodal-document-understanding-for-extraction-pipelines.json"
    },
    "next": {
      "slug": "splitting-long-documents-for-document-extraction",
      "title": "How to Split Long Documents for Extraction: Batches, Overlap and Reassembly",
      "description": "How to split long documents for extraction: size batches from measured tokens per page, choose section or page windows, overlap pages, carry table headers, reconcile conflicts across batches and reassemble with provenance.",
      "kind": "sub",
      "order": 11,
      "html": "https://changegamer.ai/articles/splitting-long-documents-for-document-extraction",
      "markdown": "https://changegamer.ai/articles/splitting-long-documents-for-document-extraction.md",
      "json": "https://changegamer.ai/api/articles/splitting-long-documents-for-document-extraction.json"
    }
  },
  "resources": [
    {
      "slug": "document-extraction-for-agents",
      "html": "https://changegamer.ai/resources/document-extraction-for-agents",
      "markdown": "https://changegamer.ai/resources/document-extraction-for-agents.md",
      "json": "https://changegamer.ai/api/resources/document-extraction-for-agents.json"
    },
    {
      "slug": "structured-outputs-and-json-mode",
      "html": "https://changegamer.ai/resources/structured-outputs-and-json-mode",
      "markdown": "https://changegamer.ai/resources/structured-outputs-and-json-mode.md",
      "json": "https://changegamer.ai/api/resources/structured-outputs-and-json-mode.json"
    },
    {
      "slug": "evaluating-ai-agents",
      "html": "https://changegamer.ai/resources/evaluating-ai-agents",
      "markdown": "https://changegamer.ai/resources/evaluating-ai-agents.md",
      "json": "https://changegamer.ai/api/resources/evaluating-ai-agents.json"
    }
  ]
}