Mixed packets slow the intake queue.
A single upload may contain forms, IDs, invoices, statements, letters, notes, and supporting evidence. Someone still has to figure out where each document begins and ends.
Chicago automatically splits mixed files, packets, scans, and attachments into distinct business documents, then classifies each one by meaning — so Nashville, Orion, Polaris, and your downstream systems start with clean, reliable input.
Before AI can extract facts, answer questions, or trigger workflows, it needs to know what it is looking at. Chicago answers the first two questions in every document workflow: where does one document end, and what type of document is it? Once Chicago separates and identifies the content, Nashville extracts the facts, Orion reasons over the evidence, and Polaris moves the work forward.
Chicago is the first layer. Without it, every other stage works from messy input — merged packets, mislabeled files, scans of unknown type. Get this step right, and every downstream step gets easier.
Organizations do not receive neat, single-purpose documents. They receive claim packets, loan files, application bundles, scanned batches, email attachments, faxes, supporting records, duplicate pages, and merged PDFs.
Before any system can extract or reason, someone has to separate the content and identify what each document is. When that first step is wrong, everything downstream gets worse.
A single upload may contain forms, IDs, invoices, statements, letters, notes, and supporting evidence. Someone still has to figure out where each document begins and ends.
Document formats change. Vendors use different layouts. Scans are noisy. Similar-looking documents may mean very different things.
The wrong split leads to missing fields. The wrong label leads to wrong routing. The wrong context leads to wrong answers and manual cleanup.
Chicago is Haystac’s document separation and classification layer. It automatically breaks mixed document batches into individual business documents and identifies each one by meaning. It does not rely only on fixed templates, brittle keyword rules, or layout matching. Chicago uses semantic and visual understanding to recognize what a document represents, even when formats vary. That means downstream systems start with the right document, in the right category, routed to the right place.
Split merged PDFs, scanned batches, faxes, uploads, and attachments into distinct business documents.
Classify content as a claim form, policy page, medical bill, ID, application, invoice, statement, contract, letter, or supporting record.
Send each document to Nashville for extraction, Orion for reasoning, Polaris for workflow action, or an external system.
Adapt to new document types, changing formats, and evolving intake streams with less manual rule maintenance.
Chicago organizes inbound content before business logic is applied. It separates, classifies, labels, and routes documents so every downstream step starts from a cleaner input.
Documents arrive from scanners, digital mailrooms, file systems, ECM platforms, workflow tools, host applications, or inbound APIs.
Chicago separates mixed batches into distinct documents, even when a single file contains multiple forms, attachments, letters, or supporting records.
Chicago creates representations of the document’s meaning, structure, and visual patterns so classification is not limited to keywords or templates.
Documents are labeled based on semantic similarity and context, even when layouts, vendors, scans, or wording differ.
Outputs can be returned to the host system, routed to Nashville for extraction, sent to Orion for reasoning, used by Polaris for workflow execution, or passed into external applications.
Rules-based systems work when documents are predictable. They look for fixed layouts, keywords, barcodes, page positions, or template matches. Chicago takes a different approach. It uses embedding-based understanding to compare documents by meaning, structure, and visual context. Two invoices can look different and still be recognized as invoices. Two denial letters can use different wording and still be grouped correctly. A new vendor format can be classified without rebuilding an entire rule library.
Chicago does not just ask, “What does this page look like?” It asks, “What is this document?”
Every downstream step depends on Chicago getting the first step right. Extraction, reasoning, routing, and automation all become more accurate when the system starts with the correct document in the correct context.
Separated, correctly labeled documents make field extraction more accurate and reduce manual correction.
When documents are organized by meaning, retrieval-grounded answers are based on the right sources.
Correct classification helps agents route, escalate, validate, and act on the right document type.
Better intake means fewer misrouted documents, fewer failed extractions, and less manual cleanup.
Automatically split mixed files, packets, scanned batches, and attachments into individual business documents.
Classify documents by meaning and context instead of relying only on keywords, file names, or fixed layouts.
Use both document appearance and content meaning to identify documents more reliably.
Group similar documents across large repositories to identify recurring document types and prepare content for downstream AI.
Return labeled, structured outputs that can be sent to host systems, Nashville, Orion, Polaris, or external workflows.
Adapt to new document types and changing intake patterns with less manual rule-building and template maintenance.
Separate claim forms, policy pages, medical bills, adjuster notes, photos, correspondence, and supporting evidence before extraction or adjudication.
Organize loan applications, IDs, income statements, bank statements, disclosures, KYC documents, and underwriting materials.
Separate applications, eligibility forms, notices, supporting records, appeals, permits, and case correspondence.
Organize prior authorization packets, referrals, medical records, lab results, plan documents, and provider correspondence.
Classify and route scanned batches, inbound PDFs, email attachments, faxes, and business correspondence at intake.
Cluster and label large volumes of historical content so archives become usable for extraction, search, reasoning, and governance.
Chicago is the first layer of OmniSuite™. It gives the rest of the platform clean, labeled, context-aware documents to work with.
Separates, classifies, and routes inbound content.
Extracts fields, tables, entities, handwriting, clauses, and relationships.
Reasons over structured data and verified enterprise content.
Routes, escalates, validates, and triggers workflow actions.
Chicago identifies the document. Nashville extracts the facts. Orion reasons from them. Polaris moves the work forward.
Chicago separates mixed inbound content into individual business documents and classifies each document by meaning, so downstream systems know what they are processing.
No. Digital mailroom is a strong fit, but Chicago also supports claims intake, loan processing, government case intake, healthcare packets, repository cleanup, and any workflow where documents arrive mixed together.
Template systems rely on fixed layouts, keywords, or rules. Chicago uses semantic and visual understanding to classify documents by what they represent, even when formats vary.
Yes. Chicago is designed to detect document boundaries inside mixed files, packets, scans, and merged PDFs.
The document can be routed to Nashville for extraction, Orion for reasoning, Polaris for workflow execution, a human review queue, or an external host system.
Yes. When documents are correctly separated and labeled at intake, extraction, retrieval, reasoning, and workflow automation all start from cleaner inputs.
Chicago turns high-volume inbound content into separated, labeled, workflow-ready documents — so every downstream AI step starts with the right context.
We enable highly regulated organizations to build, govern, and operate domain-specific models within their own infrastructure and governance frameworks.
Chicago automatically splits mixed files, packets, scans, and attachments into distinct business documents, then classifies each one by meaning — so Nashville, Orion, Polaris, and your downstream systems start with clean, reliable input.
Before AI can extract facts, answer questions, or trigger workflows, it needs to know what it is looking at. Chicago answers the first two questions in every document workflow: where does one document end, and what type of document is it? Once Chicago separates and identifies the content, Nashville extracts the facts, Orion reasons over the evidence, and Polaris moves the work forward.
Chicago is the first layer. Without it, every other stage works from messy input — merged packets, mislabeled files, scans of unknown type. Get this step right, and every downstream step gets easier.
Organizations do not receive neat, single-purpose documents. They receive claim packets, loan files, application bundles, scanned batches, email attachments, faxes, supporting records, duplicate pages, and merged PDFs.
Before any system can extract or reason, someone has to separate the content and identify what each document is. When that first step is wrong, everything downstream gets worse.
A single upload may contain forms, IDs, invoices, statements, letters, notes, and supporting evidence. Someone still has to figure out where each document begins and ends.
Document formats change. Vendors use different layouts. Scans are noisy. Similar-looking documents may mean very different things.
The wrong split leads to missing fields. The wrong label leads to wrong routing. The wrong context leads to wrong answers and manual cleanup.
Chicago is Haystac’s document separation and classification layer. It automatically breaks mixed document batches into individual business documents and identifies each one by meaning. It does not rely only on fixed templates, brittle keyword rules, or layout matching. Chicago uses semantic and visual understanding to recognize what a document represents, even when formats vary. That means downstream systems start with the right document, in the right category, routed to the right place.
Split merged PDFs, scanned batches, faxes, uploads, and attachments into distinct business documents.
Classify content as a claim form, policy page, medical bill, ID, application, invoice, statement, contract, letter, or supporting record.
Send each document to Nashville for extraction, Orion for reasoning, Polaris for workflow action, or an external system.
Adapt to new document types, changing formats, and evolving intake streams with less manual rule maintenance.
Chicago organizes inbound content before business logic is applied. It separates, classifies, labels, and routes documents so every downstream step starts from a cleaner input.
Documents arrive from scanners, digital mailrooms, file systems, ECM platforms, workflow tools, host applications, or inbound APIs.
Chicago separates mixed batches into distinct documents, even when a single file contains multiple forms, attachments, letters, or supporting records.
Chicago creates representations of the document’s meaning, structure, and visual patterns so classification is not limited to keywords or templates.
Documents are labeled based on semantic similarity and context, even when layouts, vendors, scans, or wording differ.
Outputs can be returned to the host system, routed to Nashville for extraction, sent to Orion for reasoning, used by Polaris for workflow execution, or passed into external applications.
Rules-based systems work when documents are predictable. They look for fixed layouts, keywords, barcodes, page positions, or template matches. Chicago takes a different approach. It uses embedding-based understanding to compare documents by meaning, structure, and visual context. Two invoices can look different and still be recognized as invoices. Two denial letters can use different wording and still be grouped correctly. A new vendor format can be classified without rebuilding an entire rule library.
Chicago does not just ask, “What does this page look like?” It asks, “What is this document?”
Every downstream step depends on Chicago getting the first step right. Extraction, reasoning, routing, and automation all become more accurate when the system starts with the correct document in the correct context.
Separated, correctly labeled documents make field extraction more accurate and reduce manual correction.
When documents are organized by meaning, retrieval-grounded answers are based on the right sources.
Correct classification helps agents route, escalate, validate, and act on the right document type.
Better intake means fewer misrouted documents, fewer failed extractions, and less manual cleanup.
Automatically split mixed files, packets, scanned batches, and attachments into individual business documents.
Classify documents by meaning and context instead of relying only on keywords, file names, or fixed layouts.
Use both document appearance and content meaning to identify documents more reliably.
Group similar documents across large repositories to identify recurring document types and prepare content for downstream AI.
Return labeled, structured outputs that can be sent to host systems, Nashville, Orion, Polaris, or external workflows.
Adapt to new document types and changing intake patterns with less manual rule-building and template maintenance.
Separate claim forms, policy pages, medical bills, adjuster notes, photos, correspondence, and supporting evidence before extraction or adjudication.
Organize loan applications, IDs, income statements, bank statements, disclosures, KYC documents, and underwriting materials.
Separate applications, eligibility forms, notices, supporting records, appeals, permits, and case correspondence.
Organize prior authorization packets, referrals, medical records, lab results, plan documents, and provider correspondence.
Classify and route scanned batches, inbound PDFs, email attachments, faxes, and business correspondence at intake.
Cluster and label large volumes of historical content so archives become usable for extraction, search, reasoning, and governance.
Chicago is the first layer of OmniSuite™. It gives the rest of the platform clean, labeled, context-aware documents to work with.
Separates, classifies, and routes inbound content.
Extracts fields, tables, entities, handwriting, clauses, and relationships.
Reasons over structured data and verified enterprise content.
Routes, escalates, validates, and triggers workflow actions.
Chicago identifies the document. Nashville extracts the facts. Orion reasons from them. Polaris moves the work forward.
Chicago separates mixed inbound content into individual business documents and classifies each document by meaning, so downstream systems know what they are processing.
No. Digital mailroom is a strong fit, but Chicago also supports claims intake, loan processing, government case intake, healthcare packets, repository cleanup, and any workflow where documents arrive mixed together.
Template systems rely on fixed layouts, keywords, or rules. Chicago uses semantic and visual understanding to classify documents by what they represent, even when formats vary.
Yes. Chicago is designed to detect document boundaries inside mixed files, packets, scans, and merged PDFs.
The document can be routed to Nashville for extraction, Orion for reasoning, Polaris for workflow execution, a human review queue, or an external host system.
Yes. When documents are correctly separated and labeled at intake, extraction, retrieval, reasoning, and workflow automation all start from cleaner inputs.
Chicago turns high-volume inbound content into separated, labeled, workflow-ready documents — so every downstream AI step starts with the right context.
We enable highly regulated organizations to build, govern, and operate domain-specific models within their own infrastructure and governance frameworks.