Document automation is about processing documents without or with minimal human intervention. It's nothing new - the only thing new is that technology is so advanced that even complex tasks such as classification of documents can be automated.
Automation of Document Flows
The essence of automating document flows is to capture the documents from the source, process the documents so that they become uniform, extract important information for processing the document and forward the document further. Some of the concepts and technologies we work with in such a context are:
- Capture (Scanning): Whether documents are born digitally or scanned, it is important that they are captured in a stream that ensures that their treatment is no matter where the documents came from how they came here and what format they are in .
- Recognition: Recognition software can replace manual document sorting and data entry processes with automatic document recognition and data capture.
- Automated document creation: Automatic document creation involves the production of large amounts of uniform documents automatically.
- PDF Conversion: Experience says that it may be a good idea to convert the documents into a uniform and machine-readable format before they are submitted to the software that makes the recognition.
- Robots: Robots are good at repetetive tasks, and an application of software robots, we've seen is to make the robot "patch a gap" between two systems, which do not exchange data through a technical integration, but the robot can make the data exchange through the system user interfaces.
Several types of software offer document classification abilities. The newest kid on the block in this context is artificial intelligence. We work with a product that for example can
- Go through huge amounts of files and build an index of file types, duplicates, and document text
- Classify documents based on a short training
- Extract entities from documents, such as geographic locations, person names, card og security numbers etc
And the above is what we need to classify huge amounts of unclassified documents and point out documents with Personally Identifiable Information.
We are still learning the abilities of this extraordinary technology and challenging the boundaries, but what we see is impressive. The surprising part is that products are accessible even for a modest budget.
How to get started?
You want to automate document tasks, but an automation project is huge and expensive - or is it? Is it possible to start small and expand from there? Yes, and in many cases, we see that the customer already has a lot of technology that can be leveraged.