Back to ABBYY Blog
More Flexibility, Skills, and Accuracy in the Latest ABBYY Vantage Release
Bill Galusha
June 07, 2022
New innovative capabilities drive greater process efficiencies across all business areas for automating document processes
When discussing intelligent automation with our customers, one key challenge that rises to the top of the conversation is how the data trapped in their documents is the lifeline to many of their processes. They lack the ability to automatically identify the documents, extract the data, and connect it to their employees and processes, and they’re feeling the pain both in back-office operations and on the customer experience.
We hear you loud and clear, and that is why we’re excited to announce the new capabilities of the latest Vantage 2.2 release that align to your company’s business goals and outcomes.
There’s something in this release for all users, whether you’re a non-coder (aka Citizen Developer), someone with experience in training document models, or simply the human operator whose job is to work with the data that is being processed. Request a demo today to access the leading ABBYY Vantage Intelligent Document Processing (IDP) platform and download the latest skills from the ABBYY Marketplace.
Check out the ABBYY Vantage 2.2 highlight recording
An ever-growing catalog of ready-to-go document automations
Last year, we introduced the ABBYY Marketplace, providing every user with access to a catalog of trained skills and other intelligent document process assets like connectors. In this release, we’re excited to let you know that the marketplace has now grown to over 100 assets, including 15 new document skills that jumpstart automation projects to process a variety of different document types.
All document skill automations, connectors, and other assets are built by ABBYY or certified partners so that your business can easily download and get started quickly, while at the same time further training and customizing the skills to meet your business needs.
The latest set of automations include document skills that are production ready:
- Remittance Advice – process proof of payment documents sent by a customer to a business, including capturing all header, footer, and line-item details.
- Utility Bills – extract supplier name, statement date, customer name and address to verify an individual’s residency.
- Personal Earnings Statement – extract data from payslips, including name, business, amounts, etc. Key to verifying an individual’s recent earnings and employment status.
- Bank Statement – extract data from bank statements including name, account, balance, and other fields which are key to verifying an individual’s or business’s financial assets.
- 1040 – extract data from Form 1040, including name, filing status, amounts, etc. Key to tax processing or verifying individual’s past earnings.
- Identity Documents (ID) – a specialized trained model that extracts information from over 10,000 identity documents in 248 different countries. Utilize this ID skill as part of verifying individuals’ identity or other status during client onboarding.
- Arrival Notice – automatically process arrival notice documents issued by a carrier including extracting the shipper, consignee, parties, carrier, and many other details about delivered goods and delivery terms.
Additionally, a new set of framework skills for processing 1040 (Schedule C, Schedule E, Schedule F, Schedule K-1), lease agreements, W8-BEN, and Sick Notes (German) are available in the ABBYY Marketplace for users to try out.
Introducing ID Skill to accelerate client onboarding and servicing
Every business wants to deliver a seamless online mobile experience to customers. The ability to verify the identity of an individual is becoming increasingly important. With the new Identity Document (ID) skill, ABBYY is delivering the most comprehensive trained model that extracts information from passports, US driver licenses, and many more. In fact, the new ID skill recognizes and extracts information from 10,000 different document types in 248 countries.
Using the Vantage Mobile App, users can snap a photo of the ID (front and back) and submit it along with other trailing documents for processing. This could be part of the process for opening a new account, applying for a new REAL ID in the US, or other client services.
To request a license and try out the new ID Skill, visit the ABBYY Marketplace.
More flexibility and higher accuracy with document training
In this release, we have introduced several capabilities designed to give users more control over training document skills using the Vantage Advanced Designer. The result is more advanced flexibility in how you can combine activities in the training pipeline to get the optimal extraction results from even the most complex unstructured documents.
Skill Designer users can leverage the following activities to tackle documents that have a high degree of variation and complexity like contracts, lease agreements, financial documents, and more.
Deep learning activity – skills are trained to extract field data from semi-structured documents using ABBYY’s underlying machine learning technology. It learns based on image patterns, the spatial structure of the document, field content, and surrounding labels. The user simply labels the document, trains the skill, and reviews the results. Deep learning training is complementary to the fast machine learning that Vantage uses and only requires a couple documents to train.
Scripting activity – a low-code design that allows users to add another layer of validation checks to the extracted data.
Improved accuracy for all document types and languages
At the core of what ABBYY has been building for years that thousands of enterprises use today is our optical character recognition (OCR) technology. Some might try to say it’s old, but at ABBYY, we like to say it is PROVEN.
Still, we are constantly evaluating and finding new ways to deliver better accuracy not just through OCR but in all our underlying artificial intelligence (AI) and machine learning (ML) technologies. In this release, we’ve included new handwriting capabilities to tackle many common forms and documents that still come filled out with handwritten text. This means no document passes through Vantage without being read and processed.
For businesses in Japan, you will be excited to know OCR improvements have been made to deliver you increased accuracy, giving you the tools to automate even further.
Finally, image quality is a challenge, and we see early AI vendor companies struggling to find the right solution. Through pretrained models, we’ve overcome challenges related to documents with complex backgrounds like birth certificates, IDs, or any document that embeds watermarks and other types of document marks that make it difficult to recognize the data.
Monitor and optimize document processes
If data is the lifeline to your business, then second up is knowing that operationally the systems and processes are performing optimally. To have that visibility we have introduced a new data warehouse into Vantage that stores all process transactional data, and in the future, quality analytics data will be available.
Built to connect into ABBYY Timeline, the new data warehouse provides a clear path forward to:
- Analyze and monitor process execution
- Manage quality and operational controls
- Ensure processes are meeting SLAs
- Analyze and monitor end-to-end flow of work
- Get an end-to-end view by combining document transactional process data with other key systems
Get started today!
At ABBYY we are moving fast, and we continue to focus on advancing our core AI technology and deliver the best no-code / low-code IDP platform. It’s why our IDP solutions have been recognized as a leader by Everest Group’s IDP PEAK Matrix 2022, read about it here.
If you have a trial of Vantage in the cloud today, you already have access to all these incredible capabilities. If you are just learning about Vantage now and have a project in mind, contact us today for a personalized meeting and demo, and then get started with your trial.
Bill Galusha