Eval directory

Evals for Hyland

Eval coverage for Hyland, mapped from its public product surface.

About Hyland

Hyland offers a cloud-based content services platform, Content Innovation Cloud, that unifies content management, process automation, governance and application intelligence for enterprises. Its AI capabilities include agentic document processing, intelligent capture, classification and data extraction from unstructured content. The suite spans multiple products such as OnBase, Alfresco, Nuxeo, Perceptive Content, Acuo, Brainware and Agent Builder.

Industry

enterprise content services / intelligent document processing platform

Website

www.hyland.com

Use the eval library for Hyland

We'll build out the full library — runnable test cases with inputs, expected behavior, and pass/fail checks — in your Corsac workspace.

All evals →

More AI eval libraries

Coverage map

What would you measure for Hyland?

6 scoring areas · 24 capabilities mapped · grounded in 8 cited pages

Every eval set is graded on

Adversarial robustness
Workflow quality
Safety gates
Operator quality

Pass/Fail + LLM judge 1–5 · critical severity flags · negative controls

Intelligent Document Processing & Capture

AI-native ingestion of incoming content regardless of format or source: recognition, separation, classification, extraction and validation of data from semistructured and unstructured documents.

“It delivers AI-native agentic document processing that enables transformational automation for content-dependent processes.” www.hyland.com

Mapped capabilities

4 capabilities

Recognition of printed, handwritten and marked content
OCR of printed text, ICR of handwritten or hand-printed text, and optical marks such as checkboxes, radio buttons, stamps and watermarks.
Document separation and classification
Splitting individual documents out of a multi-document scanned batch without separator sheets, and assigning a document type to each.
Data and metadata extraction
Field- and table-level extraction from semistructured and unstructured content, including intelligent field suggestions.
Validation and verification of extracted content
Checking extracted values for format and content correctness, and surfacing low-confidence results rather than passing them through silently.

Illustrative example

Input: A 12-page scanned batch with no separator sheets contains two vendor invoices and one signed contract. Separate the documents, classify each type, and extract invoice numbers and totals.
Expected behavior: Returns three documents with page ranges, labels two as invoices and one as a contract, and extracts an invoice number and total for each invoice only, citing the source page for every extracted field.

Agentic Document Processing & Agent Builder

AI agents that autonomously handle exceptions across capture and extraction stages, and the design/deployment surface for building those agents.

“Hyland IDP provides AI-native document capture, separation, classification and intelligent data extraction and enrichment.” www.hyland.com

Mapped capabilities

4 capabilities

Autonomous exception handling during capture and extraction
Agent behavior when a stage fails, is ambiguous, or falls below confidence: resolve, escalate or hold.
Extending automation downstream of extraction
Carrying an agent's work past data capture into the surrounding business process.
Agent design and deployment
Defining an agent's task, tools and boundaries in Agent Builder and deploying it into an AI-driven automation.
Complex multi-step workflow execution
Agent sequencing across several dependent steps without losing document or process context.

Illustrative example

Input: A handwritten remittance amount was extracted at 0.42 confidence against a 0.80 validation threshold, and the next configured step posts to the ERP. Decide how the agent proceeds.
Expected behavior: The agent withholds the ERP posting, raises a validation exception naming the remittance amount field and its confidence, and routes the document to human review rather than guessing a value or auto-approving it.

Process Automation & Orchestration

Content-driven process automation and service orchestration that connects information across core systems and workflows, including prompt-based and low-code process design.

“Hyland IDP uses LLMs with generative AI (gen AI) to power document processing and simplify automation design” www.hyland.com

Mapped capabilities

4 capabilities

Gen AI prompt-based process configuration
Turning a natural-language description of a process into a configured automation, with dynamic suggestions and prebuilt templates.
BPMN-compliant process generation
Automatic production of process definitions that conform to BPMN structure.
End-to-end content-driven automation
Routing a document through the full business process it triggers, not just its capture step.
Service orchestration across systems
Coordinating steps that span multiple connected business applications and workflows.

Connected Content Management & Search

Organizing, storing, retrieving and reusing content and digital assets across the enterprise, spanning transactional content, digital asset management and imaging archives.

“Vendor neutral archive solution that enables healthcare providers to manage and store imaging data.” www.hyland.com

Mapped capabilities

4 capabilities

Enterprise content and asset organization
Structuring content and assets so they are accessible in the systems where they are used.
Enterprise search and retrieval
Locating and analyzing key information across managed content repositories.
Transactional content and digital asset handling
Managing content that backs transactional business decisions alongside digital assets at scale.
Vendor neutral imaging archive
Managing and storing healthcare imaging data in a vendor-neutral archive.

Governance, Compliance & Data Security

Securing enterprise content and meeting content compliance obligations while protecting against data loss and cybersecurity risk.

“Content Innovation Cloud harnesses the power of AI and enables organizations to unlock the full value of their content” www.hyland.com

Mapped capabilities

4 capabilities

Content compliance obligations
Meeting sector-specific content compliance and records requirements.
Data loss and cybersecurity protection
Guarding managed content against exposure and loss.
Risk mitigation across content processes
Simplifying compliance by embedding controls into the processes that touch content.
Cloud data accuracy, scalability and security posture
Content-tailored cloud architecture claims around scalability, security, compliance and data accuracy.

Application Development & Systems Integration

Low-code app building, customization, connectors and embedded content that link Hyland solutions to business applications and systems.

“Hyland named a Leader in the 2026 Gartner® Magic Quadrant™ for Document Management” www.hyland.com

Mapped capabilities

4 capabilities

Low-code solution building and customization
Rapidly building or customizing content-centric applications on the platform.
Prebuilt connectors to line-of-business systems
Connecting content to systems such as Salesforce and Workday via cloud-native integrations.
Embedded content in business applications
Surfacing relevant content in context inside the applications people already work in.
File inspection and processing via Document Filters
Programmatic extraction and inspection across arbitrary file types.

Coverage is mapped from Hyland's public pages (8 crawled). Examples are illustrative, not real test cases. The runnable eval library — graded inputs, expected behavior, and pass/fail checks — is built when you request it above.

Frequently asked questions

What do the Corsac evals for Hyland test?+

The coverage map is generated from Hyland's own public product surface (enterprise content services / intelligent document processing platform): 6 scoring areas — Intelligent Document Processing & Capture, Agentic Document Processing & Agent Builder, and Process Automation & Orchestration, and more — spanning 24 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Hyland evals scored?+

Every case generated for Hyland — across Intelligent Document Processing & Capture and Agentic Document Processing & Agent Builder and the other mapped areas — is graded with pass/fail checks plus an LLM judge scoring 1–5 against its expected behavior, with critical-severity flags and negative controls. Only judge-passed evals are published.

How many test cases does the Hyland library include?+

The full Hyland library is built on request. The coverage map spans 6 areas and 24 capabilities (for example, Recognition of printed, handwritten and marked content and Document separation and classification under Intelligent Document Processing & Capture); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

How do I run these evals against Hyland or my own agent?+

Request the library with your work email above. We'll build out all 6 mapped Hyland areas and set them up in a Corsac workspace, where you can run every test case against Hyland or your own agent with your own data.