Production auto learning: automating intelligent data capture

    Meet Production Auto Learning, an easier way to automate your document processing, powered by artificial intelligence and machine learning.

    Subscribe to our blog

    Ever since I joined the Amplexor data capture team, I have been learning new things every day about data capture and process automation. A good example is the OpenText Intelligent Capture (former OpenText Captiva) solution.

    With data capture technologies you can quickly transform high volumes of documents into actionable data for use in vital processes and applications. It’s true, creating advanced capture processes takes time, but the automation gained justifies the investment. In most implementations, automating document recognition and classification generates a return on investment within 12 months!

    Get to know your PAL

    Many global organizations are already using these data capture technologies to automate processes and increase efficiencies, but through our experience, we’ve noticed that many companies aren’t taking full advantage of all the available capabilities. A good example of this is the Production Auto Learning (PAL) feature in OpenText Intelligent Capture.

    Powered by artificial intelligence (AI) and machine learning (ML), this module enables your system to learn about document properties and process these automatically, so you can take advantage of document capture benefits faster than before.

    Imagine your input management technology learning how to recognize, classify and distribute invoices, loan documents, among others, automatically instead of having to rely on developers or administrators to create templates. The automation of template creation reduces configuration times up to 90%.

    Teach your PAL wisely

    PAL learns through user interaction. As the amount of processed data increases, the more PAL learns and improves. So, the sooner the Production Auto Learning module is implemented during the process, the sooner it begins gathering information and delivering better results.

    As mentioned above, this intelligent capture feature is a form of supervised learning. This means the system receives a tagged sample set so that the desired outcome is known. This way, the system can learn and improve itself. It applies past knowledge to new, unidentified input.

    Even before PAL has received the appropriate amount of information needed to excel in its task, it already reduces the time spent by employees manually validating document properties.

    Why your company should consider PAL

    Companies that receive forms, applications, invoices, loan documents, policy claims and other document types benefit from this the most. It comes down to two main points:

    1. A reduction of template configuration time up to 90%.
    2. Up to 20% higher accuracy when processing various document types.

    By using a text-based classification method with semi-structured documents and files coming from various channels, every set of words that appear in the same place in different documents are identified. Project administrators no longer need to specify keywords. The software determines these automatically before classifying a document, thus decreasing the time and cost involved compared to doing these operations manually (up to 90%).

    While image analysis works best for structured forms, it’s sensitive to stains, skewing of a document, discoloration and other factors. On its own, image classification won’t give you the best results, as it’s impacted by document quality issues.

    This is why PAL combines automated image and text-based classification. Using text-based classification to find a set of “words” that appear in the same place in two or more documents, PAL provides organizations with up to 20 percent higher accuracy processing different document types, rather than manually creating rules and templates.

    In conclusion, production auto learning technology accelerates data extraction and document classification through process automation. As a bonus, data quality is also improved, by reducing errors that can arise from manual data entry. In short, you get higher data quality and a future-proof capture solution with less development and higher accuracy.

    Still unsure about whether PAL could be a feature to explore for your input management ecosystem? Send us a message through the comments below, or schedule a demo with our team to learn more about cognitive capture and look at what a typical implementation looks like, step-by-step.

    Published on    Last updated on 29/11/2019

    #Digital Transformation, #OpenText, #Data Capture

    About the author

    Jerry Rosenau is a Junior Business Consultant at Amplexor, based in Eindhoven, The Netherlands. Having had an internship in the content capture team after college graduation, Jerry now focuses on business process analysis and supporting the sales team on advising customers about the best possible capture solutions.

    SUBSCRIBE TO OUR BLOG

    Participate in this discussion