PLANET AI

Meet IDA

Meet IDA

PLANET AI

The Power of Intelligence

Intelligent Document Analysis

Intelligent Document Analysis

The Power of Intelligence

IDP Platform - IDA

REDEFINING ADAPTIVE DOCUMENT PROCESSING

REDEFINING ADAPTIVE DOCUMENT PROCESSING

With our IDP Platform – IDA –, you can capture and classify all your incoming documents, extract and validate information to increase productivity, accuracy, and efficiency.

Performance monitoring, workflow configuration, and training are available through the web interface.

IDP PLATFORM – IDA – HOW IT WORKS

IDP Platform

– IDA –

How it works

IDA’s Features

Our modular approach allows us to train and adapt all features individually.

Recognition

With IDA, 97% of scanned text is captured and read accurately. The OCR approach includes handwritten and machine-printed texts as well as a variety of languages and scripts and implementing new requirements is easy.

How does it work?

Text from incoming documents (PDF, TIFF, JPEG, etc.) is transcribed and output as a PDF or a machine-readable JSON file.

Impressive accuracy:

  • > 99 % machine-print

  • > 95 % handwritten

  • High throughput processing (GPU support)

  • Comprehensive use of Machine Learning

Recognition

Classification

Classification

By automating classification, you save time and money by eliminating manual corrections rules. Increase automation rates up to over 90% with a fully automated document workflow.

How does it work?

Single pages or entire documents can be categorized and sorted with IDA Classification after a short set-up time. Receive incoming documents and categorize them with an average accuracy of 94%. With our document splitting feature, you can separate large consecutive documents – either with a trainable AI model or rule-based.

Data-driven, based on Machine Learning:

  • Pre-trained models

  • 2 methods: Document and Page Classification

Extraction

With IDA, you can extract data from all your documents, so that you can make better and faster decisions. You can say goodbye to rule-based approaches that are error-prone and time-consuming. Get ready to discover a “no code” solution that is purely data-driven.

How does it work?

To create a fully customized extraction module, you only need to provide a few sample documents. You can efficiently extract data from text, forms, and complex documents like invoices using the Extraction Assistant (web client).

  • Data-driven, based on Machine Learning

  • Individually trainable modules

  • Various use cases – (un)structured text, tables, and more

Extraction

Detection

Detection

Barcodes, 2D codes, and signatures can all be read by IDA. Detection can be used as a prerequisite step, e. g. for a following document classification or validation.

How does it work?

We use our state-of-the-art object detection capability. Thanks to our trainable approach, we can easily implement new requirements.

Validation

A plausibility check is offered by IDA to ensure that results are as accurate as possible. A database or pre-set rules can be used to compare results from IDA Classification or Extraction.

How does it work?

The default module offers validation rules for common use cases, such as dates or other number formats. By configuring the interface, you can easily set up connections to your master data, e. g. a database.

Validation

IDA’s Add-ons

While adding functions to IDA, add-ons depend on human input.

Search

AWARD-WINNING FULL-TEXT SEARCH ENGINE

Making your digital archives searchable is a vital step toward a digital conversion process. Manual capture from paper-based input is costly, error-prone, and inefficient.

How does it work?

IDA Search leverages our patented core technology to enable full-text search in digital document archives with unprecedented accuracy.

Documents captured by IDA Recognition can be narrowed down using regular expressions, operators, and filters. Even scattered input can be recognized with the highest accuracy.

  • Internationally-awarded keyword-spotting

  • Robust and fast: hybrid search for machine-print

  • Multiple use cases: search in handwritten and machine-printed documents

Revision Assistant beta

Revision Assistant

MANUAL CORRECTION AND VALIDATION

Companies are automating their input management to the maximum extent possible. However, humans must often be involved in the process to validate or correct results that did not pass through automatic processing.

How does it work?

The IDA Revision Assistant is the next step towards a 100 % accurate document workflow. It provides a web interface to manually correct or validate results from IDA Classification and Extraction. Users can correct errors before they enter the next module – improving the accuracy and efficiency of the entire workflow.

  • Handle files with a web client made for high-throughput processing

  • Assign users to groups for different expert levels and technologies (e.g. page classification or data extraction)

  • Send reviewed documents back to your workflow, boosting overall accuracy close to 100 %

Partners

WHAT OUR CUSTOMERS AND PARTNERS SAY

Frequently Asked Questions

Sound familiar?

  • My growing document volumes stall business processes that are dependent on data
  • My inflexible software solutions have little scope for customization
  • My documents are filled with scrawly handwriting and typescript making it impossible to scan
  • My manual processes are Error-prone and too costly to capture and sort documents at scale

IDA is a software suite for intelligent document processing. Many industries can benefit from automating document-based processes with IDA. Data extraction, validation, and full-text search are among our features that cover the entire document workflow.

IDA is not a “one size fits all” solution. On our IDA Platform, we will provide software modules for standard use cases as well as trainable AI assistants for individual customization. Our “no code” approach guarantees fast and easy set-up – for an instant return on investment. We leverage patented core technology for data capture, including our award-winning handwriting recognition.

Research and lively exchange with customers and partners drive PLANET AI’s product development. This makes automation with IDA a valuable investment for the future of your business.

IDA can process various PDF and image formats as well as JSON input, depending on your feature and add-on selection.

Documents may contain handwritten and typescript content in different languages and scripts. With our trainable approach, IDA can adapt to any use case or challenge.

With PLANET AI, you can access an easy-to-use software platform complete with tutorials, online documentation, and help desk support.

The development of PLANET AI products is based on state-of-the-art research and active customer and partner interaction.

Therefore, IDA is always evolving to become a platform with deep content understanding.

Do you have a company in mind that might benefit from AI automation?

As part of our implementation services, we consult several integration partners on how to implement IDA on the customer’s site. In most cases, we will start with a Proof of Concept (or MVP) before going live.

Is IDA something you would like to incorporate into your business workflow? For IDA integration, we can either provide you with an appropriate integration partner or train your specialists.

It would be our pleasure to consult with you! Go to our Contact Us page to take the next step in automation!