Cognitive document processing makes organizations faster

Publish date:

Processing large volume of paper documents requires a lot of manual efforts. Our new solution Cognitive Document Processing can reverse this trend through automation. Read further to learn more.

Input management – many people don’t know what it is. In short, input management refers to the systems that ensure that information from outside an organization arrives in an orderly manner and is distributed through the organization. In the past, this involved paper documents delivered by post or by fax. The method most large organizations have started involves digitizing documents in a central mail room. An electronic document is created from the paper pages by scanning it. After scanning, this document is classified and distributed to the employees who have to process it.

Electronic forms

Because the processing of paper documents requires a lot of manual work, many organizations have switched to electronic document and messages. The information is input by means of electronic forms. Because the data is already machine-readable on the form, the elaborate digitization and correction of data are no longer necessary. Many organizations have encouraged, and sometimes even prescribed, the use of electronic forms. For example, the Dutch social security institute UWV only accepts, in principle, electronic forms and other documents. Paper is ruled out.

The paper mail flow is steadily declining. As a side-effect, central mailrooms, where all written communication arrives, are also disappearing. Messages, such as emails, are sent directly to (groups of) employees. In this way, organizations can lose their grip on communication flows. This implies that information gets lost, messages are forgotten or handled incorrectly, or unauthorized persons can gain access to confidential information.

Social media

The number of ways in which we communicate is also growing. In addition to email, a great many social media channels have been added as a communication channel for organizations. Customers can send questions and messages via Facebook, Twitter, Google, Instagram, and so on. All these channels must be monitored and the important issues detected and processed accordingly. Currently, this is mostly done manually.

And so, we are shooting ourselves in the foot again. We thought we could streamline incoming communication through the use of electronic forms, but the proliferation of social media has increased the number of communication channels. By using structured forms, we thought we could identify data elements unambiguously, the “family name” field records the family name, but in social media, the messages are free format. We may have to read the entire message to find the family name if it is mentioned at all.

Cognitive document processing

Fortunately, there are solutions that can reverse this trend, and they involve artificial intelligence (AI). With AI, free text can be read and interpreted. This form of AI, natural language processing (NLP), makes it possible to recognize the relevant concepts in a text and identify and classify the message based on those concepts. Classification is already possible based on words and the place of words in the text. But NLP can also place words in their context in the sentence or sentences. We call this a concept extraction. This way, we not only read the word “complaint,” but also determine what the complaint is about. For example, there is a substantial difference between the phrase “I have a complaint about your ACME product” and the phrase “Your handling of my complaint has been perfect.” Cognitive computing with machine learning can recognize such differences. I am not saying that this is going to be easy, but by properly training the system, we can achieve results quickly.

Automatic analysis of texts has two uses that can already be delivered with the current state of AI. First, classification of the message or document contents: this technique allows the message to be archived in the right place, with the associated retention periods and, possibly, the correct security level. Secondly, starting the right business process to handle the message: in addition to these operational use cases, we can also use NLP for data analysis – for trends and forecasts — and knowledge management. Many articles have been published on this topic recently, but my advice is: if you want to get benefits fast, start with a more efficient implementation of the document-based bulk processes of your organization.

A lot of work can be saved with AI, such as Capgemini’s Cognitive Document Processing. By automating cognitive skills in input management, organizations can process the flood of messages from each channel in a structured way. And the processing of messages and documents can go faster. So, in turn, organizations can respond more quickly to the needs of their customers. In result, your business becomes better than the competition.

Don’t assume the ways you handle your incoming communication is efficient and adequate. Modern technology offers a lot of possibilities to automate and streamline input management. My advice is, take a new look at your processes around incoming communication and start making these processes better, stronger, faster.

Related Posts


Validation techniques to improve your data quality

Stephen Timbers
Date icon April 4, 2019

Wish to improve your data quality? Read further to learn how that can be achieved.


If it’s not instant – It’s not good enough

Vidur Thakral
Date icon March 25, 2019

Are your technical processes keeping up the pace with the digital boom?


By continuing to navigate on this website, you accept the use of cookies.

For more information and to change the setting of cookies on your computer, please read our Privacy Policy.


Close cookie information