When it comes to process automation, of the most common, persistent workflow bottlenecks companies of all sizes deal with is manual document handling and data entry. Despite decades of investments in technology and automation with the goal of freeing workers from their keyboards, the reality is that most common automated business processes still depend on manual handling of documents and data entry.
Of course, that model is changing as companies seek to embrace digital transformation and adopt technologies that include an intelligent data capture component, like Artsyl’s docAlpha. The docAlpha transformation platform does far more than classify documents and extract process-relevant information. It can also intelligently route documents for coding, approval or matching, or easily integrate with ERP or ECM systems that facilitate workflow and create transactions/upload documents into those systems.
But, at is core, docAlpha is a transformation engine—software that converts unstructured content like invoices, orders, quotes or receipts that show up as unstructured Word docs, PDFs or paper, into specific, structured data that fits neatly into the structured data fields within an ERP system or a document management system.
How well any intelligent document capture system can accomplish this task is one of the keys to its success. The time it takes to correctly identify the right data, without repeated human intervention, impacts the overall speed, performance and effectiveness of any document-dependent process.
That’s why Artsyl continues to push the envelope, in each product release, in terms of the speed, performance and flexibility of its intelligent data capture capabilities.
Often, one of the biggest opportunities for overall performance improvement lies in the ‘Auto-find’ capabilities of docAlpha. This is the component that is able to ‘read’ a document to identify key data elements like an invoice number, a vendor or customer name, amount subtotals and totals, etc.
In fact, the release of docAlpha continued an overall trend of increased Auto-find performance. The latest docAlpha enhancements increase recognition speeds by 180% compared to when docAlpha was first released late in 2018.
Keep in mind that the recognition throughput for Auto-find in that initial version of docAlpha was 210% faster than docAlpha.By 6.0, Artsl rebranded its “Auto-Find” function as “Advanced Auto-Find” because it introduced a knowledgebase for finding data that could take priority over pre-configured document definitions, along with user-friendly, visual, mouse-driven modifications that could be made to the location of data. The latest Artsyl release built upon these concepts even further, providing more contextual guidance for how the system can find relevant data. For example, docAlpha may key off a text “anchor” word like “Total” and then know to look nearby to find a dollar amount. This improves BOTH accuracy and speed.
Beyond pure performance improvements, applying machine learning to Auto-find and document capture has reduced the time and effort required to set up and deploy docAlpha. Today, machine learning allows users to “train” docAlpha whenever an exception requires human intervention. By allowing docAlpha to record user interaction and then add those new instructions to its knowledgebase, the system learns and improves continuously.
Creating a system that can learn and adapt and overcome reliance on more rigid, programmatic approaches means less human intervention overall, and reduces dependence on IT or other technical resources.
Instead, the power is in the hands of the process owner.
For more information about Intelligent Process Automation and data capture, contact your Artsyl Technologies Account Representative.