Turn documents into trusted business data with ease

Add AI-powered classification, extraction, and document understanding when traditional rules and templates reach their limits. IRISPulse AI Xtract processes complex layouts, handwriting, tables, and unstructured content, delivering accurate, structured data to the applications and workflows your organization already uses. With a fast, no-code setup and intuitive validation experience, teams can accelerate automation and achieve better extraction outcomes with less effort.

Turn documents into trusted business data with ease

Broader automation

Handle document cases that are difficult for fixed templates.

Less manual effort

Reduce repetitive interpretation and data entry.

Protect existing investment

Extend supported IRIS environments without replacing them.

More controlled AI

Manage services through monitoring, lifecycle and governance.

Apply AI where document complexity creates work

IRISPulse AI Xtract is the AI core of IRISPulse. It supports IRISPulse applications, partner verticals and supported on-premises IRIS products with shared services for classification, extraction and document understanding.

Understand more document types

Understand more document types

Use AI services for difficult layouts, unstructured content, handwriting and complex tables.
Return usable results

Return usable results

Transform document content into structured information ready for validation, workflow and integration.
Extend, do not replace

Extend, do not replace

Add cloud AI and LLM capabilities to supported IRISXtract environments while protecting existing rules, projects and integrations.

From client application to structured result

Step Icon

Client application

An IRISPulse application, IRIS product or external application submits the document.

Step Icon

IRISPulse

A workflow triggers the selected IRISPulse AI Xtract service.

Step Icon

AI service

The selected extraction or classification service runs.

Step Icon

AI processing

AI extracts the required information from the document.

Step Icon

Structured result

The result is ready for validation, routing or integration.

Invoices and purchase orders

Invoices and purchase orders

Extract business information from transactional documents for downstream processes.

Delivery notes

Delivery notes

Structure relevant information from operational documents.

Contracts

Contracts

Apply document understanding to content that does not follow a simple fixed layout.

ID cards

ID cards

Use a pre-built extraction service for identity document content.

Handwriting

Handwriting

Extend automation to handwritten information using shared AI capabilities.

Complex tables

Complex tables

Extract structured information from table-based document content.

Select the right model approach for each service

IRISPulse AI Xtract combines a service catalogue with a model layer designed for multiple AI model providers and deployments. IRIS is choosing the ones that best fit your needs; Having the right AI for the right Job. The portfolio includes open-source LLMs, IRIS-optimized models, customer-specific fine-tuned models and an AI Service Designer for partners and customers.

  • Compose services
    Create document extraction, classification, enrichment or entity-recognition services.
  • Configure models and prompts
    Select and configure models, define prompts and publish services to the catalog.
  • Operate through a shared runtime
    Use execution, orchestration, monitoring, telemetry, security and governance services.

IRISPulse AI Xtract Service Architecture

  • AI Service Catalog

    Extraction · classification · enrichment · entity recognition

  • Vendor-agnostic model layer

    Open source · IRIS models · customer-specific models

  • Platform and runtime

    Execution · routing · load balancing · monitoring · telemetry

  • Security and governance

    Access control · audit · quality oversight

Manage AI services as controlled business capabilities

The IRISPulse AI Xtract lifecycle includes service states, versioning, controlled rollout, compatibility management, deprecation warnings and platform-level management.

Design

Design

Define and prepare an AI service.
Ready

Ready

Make the service available for usage.
Deprecated

Deprecated

Retain access while signaling that replacement is recommended.
Archived

Archived

Retire the service when it is no longer available.

Start with the documents that create the most exceptions.

Show us the formats, layouts, handwriting, tables and extraction challenges that limit your current process.