{"observation":{"id":"d1acf6ae-95ec-460a-bac2-aed663d491c5","tool":"nanonets","tool_name":"Nanonets","criterion":"schema-adherence","criterion_name":"Schema Adherence","criterion_definition":"Does the output follow the supplied JSON schema hierarchy exactly, with correct nesting, field names, and data types?","criterion_evidence_type":"transformation","criterion_rank_role":"decisive","criterion_rank_role_reason":"If the output does not match the requested JSON schema exactly, the extracted data cannot be reliably consumed or queried, so this is core to the task. (3 of 3 judges)","scenario":"invoice-pdf","scenario_name":"Invoice PDF","group_tag":"financial-document-extraction","scenario_description":"A 2-page broadcast advertising invoice PDF with 8 line items, complex time/day fields, large dollar amounts, and compliance text, used to test hierarchical line-item extraction and financial validation.","modality":"pdf","input_text":null,"input_artifact_refs":[{"alt":null,"url":null,"role":null,"filename":"Invoice PDF.pdf"}],"stresses":["Nested line-item hierarchy extraction","Multi-page line-item continuity across a page break","Large dollar amount precision and total validation","Parsing time slots, day patterns, and air dates","Extraction of alphanumeric ad IDs and reference codes","Structured metadata mapping for advertiser, station, billing, and remit sections","Political advertising and FCC compliance text recognition"],"verdict":"worked","score":null,"score_total":null,"note":"It preserves the invoice hierarchy exactly as requested, filling dedicated objects for invoice_metadata, advertiser, station, account_details, billing_address, remit_address, flight_dates, line_items, and summary.","evidence_state":"verified","source":null,"artifacts":[{"url":"https://d3epheqghktydj.cloudfront.net/research-media-docstrange-nanonets-invoice-output-9dc0f476da4c.json","role":"output","alt":null},{"url":"https://cdn.futuresmart.ai/public/aidemos/aee556df2fac4352897e803fbe332c53.mp4?v=1","role":"context","alt":null}],"run_id":"ec4d736d-95f9-4c88-884c-e280435f7b7b","study_title":"Extract and query structured data from documents using natural language","study_kind":"generation","research_task":"86b9y25e5","tested_at":null,"completeness":"output-only","input":{"state":"not-captured","text":null,"files":[],"modality":"pdf","stresses":["Nested line-item hierarchy extraction","Multi-page line-item continuity across a page break","Large dollar amount precision and total validation","Parsing time slots, day patterns, and air dates","Extraction of alphanumeric ad IDs and reference codes","Structured metadata mapping for advertiser, station, billing, and remit sections","Political advertising and FCC compliance text recognition"]},"tool_page_slug":"nanonets","tool_url":"https://aidemos.com/tools/nanonets","permalink":"https://aidemos.com/evidence/d1acf6ae-95ec-460a-bac2-aed663d491c5","api_url":"https://ai.aidemos.com/v1/observations/d1acf6ae-95ec-460a-bac2-aed663d491c5"},"peers":[{"id":"c01474f8-d601-4535-b19c-9ab5759d7f35","tool":"datalab","tool_name":"Datalab","verdict":"worked","score":null,"score_total":null,"note":"Reconstructs the invoice into nested JSON objects and arrays that follow the supplied schema, including invoice_metadata and line_items rather than returning generic OCR text.","artifact_count":5,"thumbnail":"https://d3epheqghktydj.cloudfront.net/research-media-datalab-invoice-metadat-1c39e0c1c45e.png","evidence_url":"https://aidemos.com/evidence/c01474f8-d601-4535-b19c-9ab5759d7f35"},{"id":"ad656fd6-a7c6-4174-a1a5-71f32b26d964","tool":"extend-ai","tool_name":"Extend AI","verdict":"worked","score":null,"score_total":null,"note":"The invoice output is reconstructed as nested JSON sections — invoice_metadata, advertiser, station, addresses, flight_dates, line_items, and summary — instead of raw OCR.","artifact_count":2,"thumbnail":"https://d3epheqghktydj.cloudfront.net/research-media-extendai-invoice-output-9eaac6626f6c.json","evidence_url":"https://aidemos.com/evidence/ad656fd6-a7c6-4174-a1a5-71f32b26d964"},{"id":"e21e3065-67a9-44b0-a7d5-4534cc545780","tool":"landing-ai","tool_name":"Landing AI","verdict":"worked","score":null,"score_total":null,"note":"Reconstructs the requested invoice hierarchy in structured JSON, filling invoice_metadata, advertiser, station, line_items, and summary objects instead of returning a flat extraction.","artifact_count":4,"thumbnail":"https://d3epheqghktydj.cloudfront.net/research-media-landing-ai-extracted-invoice-metadata-57a291c39e3d.png","evidence_url":"https://aidemos.com/evidence/e21e3065-67a9-44b0-a7d5-4534cc545780"},{"id":"0b3ba3e3-dd88-4377-97f7-2da41c08c915","tool":"llamaparse","tool_name":"LlamaParse","verdict":"worked","score":null,"score_total":null,"note":"Keeps the invoice hierarchy intact, emitting invoice_metadata, advertiser, station, line_items, and summary objects rather than flat OCR text.","artifact_count":2,"thumbnail":"https://d3epheqghktydj.cloudfront.net/research-media-llamaparse-invoice-metadata-af804b493598.png","evidence_url":"https://aidemos.com/evidence/0b3ba3e3-dd88-4377-97f7-2da41c08c915"},{"id":"4d495002-374a-4396-8f2b-a4ff9d296493","tool":"reducto","tool_name":"Reducto","verdict":"mixed","score":null,"score_total":null,"note":"Leaks schema-unrequested fields into the invoice export, including idb_number, so the output is not strictly constrained to the provided schema.","artifact_count":3,"thumbnail":"https://cdn.futuresmart.ai/public/aidemos/752480041e324ae5a3b8f36680d83479.png?v=1","evidence_url":"https://aidemos.com/evidence/4d495002-374a-4396-8f2b-a4ff9d296493"},{"id":"38fc2de5-4406-486b-9a50-fa39c4b5e88a","tool":"retab","tool_name":"Retab","verdict":"worked","score":null,"score_total":null,"note":"Builds the requested invoice hierarchy with separate invoice_metadata, advertiser, station, account_details, billing_address, remit_address, flight_dates, line_items, and summary objects instead of a flat OCR dump.","artifact_count":3,"thumbnail":"https://d3epheqghktydj.cloudfront.net/research-media-retab-extracted-invoice-metadata-2b6610b8f928.png","evidence_url":"https://aidemos.com/evidence/38fc2de5-4406-486b-9a50-fa39c4b5e88a"},{"id":"15ef12f7-5460-42d4-99d5-9ee4785c8552","tool":"unstract","tool_name":"Unstract","verdict":"worked","score":null,"score_total":null,"note":"Populates the requested invoice hierarchy into dedicated sections for invoice_metadata, advertiser, station, account_details, billing_address, remit_address, flight_dates, line_items, and summary.","artifact_count":2,"thumbnail":"https://cdn.futuresmart.ai/public/aidemos/18ce6c9c9f314db694ae8439172b4041.png?v=1","evidence_url":"https://aidemos.com/evidence/15ef12f7-5460-42d4-99d5-9ee4785c8552"}],"other_criteria":[{"id":"0fdd65d7-8e67-48e3-94c1-faf11b60bf26","criterion":"extraction-accuracy","criterion_name":"Extraction Accuracy","rank_role":"decisive","verdict":"worked","score":null,"score_total":null,"note":"It extracts invoice metadata and financial totals accurately, including 8 aired spots, gross_total 29750, agency_commission 4462.5, net_amount_due 25287.5, and payment_terms \"30 Days\".","artifact_count":3,"evidence_url":"https://aidemos.com/evidence/0fdd65d7-8e67-48e3-94c1-faf11b60bf26"},{"id":"47a6fef2-07cd-4dea-97ff-47ef7be54f95","criterion":"extraction-accuracy","criterion_name":"Extraction Accuracy","rank_role":"decisive","verdict":"mixed","score":null,"score_total":null,"note":"It concatenates program-description text without a separator, so labels such as \"Political Issue Rates\" are appended directly onto the show title in fields like \"Fox 9 AM News at 4:30amPolitical Issue Rates\".","artifact_count":3,"evidence_url":"https://aidemos.com/evidence/47a6fef2-07cd-4dea-97ff-47ef7be54f95"},{"id":"c3d6a0fb-4f54-4ee8-9e7a-c8382539a82e","criterion":"semantic-field-enrichment","criterion_name":"Semantic Field Enrichment","rank_role":"decisive","verdict":"mixed","score":null,"score_total":null,"note":"It leaves line 8 without derived scheduling fields such as flight_period_start, flight_period_end, frequency, and days_pattern, even though comparable rows populate them.","artifact_count":2,"evidence_url":"https://aidemos.com/evidence/c3d6a0fb-4f54-4ee8-9e7a-c8382539a82e"},{"id":"7cf62e7e-87da-4954-bf96-84c30a5465e3","criterion":"semantic-field-enrichment","criterion_name":"Semantic Field Enrichment","rank_role":"decisive","verdict":"worked","score":null,"score_total":null,"note":"The tool correctly enriches ordinary invoice rows with derived scheduling fields such as time_slot, day_of_week, flight_period_start/end, frequency, and days_pattern on line items like line 4.","artifact_count":2,"evidence_url":"https://aidemos.com/evidence/7cf62e7e-87da-4954-bf96-84c30a5465e3"},{"id":"513d08f7-b8e4-41ad-ad57-d1d53f2bd1ed","criterion":"table-and-record-completeness","criterion_name":"Table & Record Completeness","rank_role":"decisive","verdict":"worked","score":null,"score_total":null,"note":"It extracts all 8 advertising line items as separate records without row merging or duplication.","artifact_count":1,"evidence_url":"https://aidemos.com/evidence/513d08f7-b8e4-41ad-ad57-d1d53f2bd1ed"}],"appears_in":[{"page_type":"ranking","slug":"document-extraction","title":"Best AI Tools for Extracting Structured Data from PDFs and Business Documents","url":"https://aidemos.com/best/document-extraction","binding":"run"},{"page_type":"tool","slug":"unstract","title":null,"url":"https://aidemos.com/tools/unstract","binding":"run"}],"same_scenario":[{"id":"c01474f8-d601-4535-b19c-9ab5759d7f35","tool":"datalab","tool_name":"Datalab","verdict":"worked","score":null,"score_total":null,"note":"Reconstructs the invoice into nested JSON objects and arrays that follow the supplied schema, including invoice_metadata and line_items rather than returning generic OCR text."},{"id":"ad656fd6-a7c6-4174-a1a5-71f32b26d964","tool":"extend-ai","tool_name":"Extend AI","verdict":"worked","score":null,"score_total":null,"note":"The invoice output is reconstructed as nested JSON sections — invoice_metadata, advertiser, station, addresses, flight_dates, line_items, and summary — instead of raw OCR."},{"id":"e21e3065-67a9-44b0-a7d5-4534cc545780","tool":"landing-ai","tool_name":"Landing AI","verdict":"worked","score":null,"score_total":null,"note":"Reconstructs the requested invoice hierarchy in structured JSON, filling invoice_metadata, advertiser, station, line_items, and summary objects instead of returning a flat extraction."},{"id":"0b3ba3e3-dd88-4377-97f7-2da41c08c915","tool":"llamaparse","tool_name":"LlamaParse","verdict":"worked","score":null,"score_total":null,"note":"Keeps the invoice hierarchy intact, emitting invoice_metadata, advertiser, station, line_items, and summary objects rather than flat OCR text."},{"id":"4d495002-374a-4396-8f2b-a4ff9d296493","tool":"reducto","tool_name":"Reducto","verdict":"mixed","score":null,"score_total":null,"note":"Leaks schema-unrequested fields into the invoice export, including idb_number, so the output is not strictly constrained to the provided schema."},{"id":"38fc2de5-4406-486b-9a50-fa39c4b5e88a","tool":"retab","tool_name":"Retab","verdict":"worked","score":null,"score_total":null,"note":"Builds the requested invoice hierarchy with separate invoice_metadata, advertiser, station, account_details, billing_address, remit_address, flight_dates, line_items, and summary objects instead of a flat OCR dump."},{"id":"15ef12f7-5460-42d4-99d5-9ee4785c8552","tool":"unstract","tool_name":"Unstract","verdict":"worked","score":null,"score_total":null,"note":"Populates the requested invoice hierarchy into dedicated sections for invoice_metadata, advertiser, station, account_details, billing_address, remit_address, flight_dates, line_items, and summary."}]}