Overview
At Appian, we know AI works best inside a process. As you see the power of AI in processes involving documents, you may want to know how to expand these capabilities across your organization.
Introducing AI Document Center: your one-stop-shop application for enterprise-grade intelligent document processing. Quickly define your document types or fields to extract in one intuitive workflow, and let Appian create the underlying design objects for you. AI Document Center is up to the challenge of handling your business’s highly complex documents with varied formatting and layouts, leading to higher classification and extraction accuracy.
AI Document Center allows users, of all technical skill levels, to create and refine models for both document classification and extraction. It supports versioning of models and has a robust reconciliation and accuracy metrics framework giving your users the tools they need to take full advantage of advances in AI capabilities. AI Document Center will generate (and update) the AI Skill objects and their associated process models that can be implemented in any application.
Key Features & Functionality
Extraction Development and Monitoring - Supports structured and unstructured extraction models. Fields can be configured manually or using an extraction tool. Models are versioned and extraction accuracy at the field level is monitored. Reconciliation is supported.
Classification Monitoring - Supports generative and ML-based classification models. Models are versioned and classification accuracy is monitored. Reconciliation is supported.
Production Metrics: The application includes a metrics framework that allows for tracking performance of AI models not just in testing but also at scale when used in production applications.
Monitoring model performance over time: By versioning models and tracking execution statistics like accuracy, the application provides visibility into how well models are performing and where improvements may be needed. This helps ensure models remain effective as data and use cases evolve.
Standardizing model deployment: The application aims to standardize how models are configured, versioned, tested and deployed across environments. This helps improve reproducibility, as well as governance over how AI/ML solutions are managed within the business.
Reconciliation - Framework and custom reconcile forms are supported.