Cognitive Automation

Osmond Li, Senior IT Manager of Technology Innovation at Dah Chong Hong (DCH) Holdings Ltd

Osmond Li, Senior IT Manager of Technology Innovation at Dah Chong Hong (DCH) Holdings Ltd

Osmond Li, Senior IT Manager of Technology Innovation at Dah Chong Hong (DCH) Holdings Ltd., has many years of enterprise IT experience including Robotic Process Automation (RPA), Workflow Automation, Machine Learning, Data Warehouse, ERP, etc. DCH began looking into new technologies such as AI/ML and RPA a few years for a companywide transformation to modernize existing processes and reduce operational costs.

After conducting an introductory RPA workshop, the first RPA bot was put into production in just one month. By the second month, five RPA bots were in production including a cognitive RPA, i.e. a bot able to learn new invoice format through optical character recognition (OCR) and machine learning.

What is Cognitive Automation?

Cognitive automation (or Cognitive RPA) is one kind of automation by combining artificial intelligence (such as natural language processing, text analytics, and machine learning), and process automation capabilities to improve business productivity and efficiency, conform to compliance and quality, and reduce turnaround and error rates. It is the process of digitizing, augmenting and automating enterprise-wide decision-making processes. The main benefit of cognitive automation is that it helps weave unstructured data from documents, customer interactions, voice, and machine vision into business workflows.

RPA, robotic process automation, is a software to automate repetitive tasks within business and IT processes via software scripts that emulate human interaction with the application user interface. Other than this, RPA can provide screen scraping and scripting programming. The key benefit of RPA is that it can take up the labour-intensive tasks with structured data in a well-defined rule.

Thanks to the blooming of AI/ML, RPA has evolved to integrate with these advanced technologies to provide Cognitive RPA. This integration helps organizations extend automation to business processes, making the most of not only structured data, but also growing volumes of unstructured information. Unstructured information such as customer interactions, unstructured documents (like invoices from different suppliers) can be easily analyzed, processed, and structured into useful data for the next steps of the process, such as predictive analytics, preventive decision making, and responsive chatting.

Beauty of Cognitive Automation

The key advantage of automation is to offload employees from high-volume repetitive tasks which employees may need to work overtime or on weekends. By automating these, a substantial amount of manpower can be saved and employees can take up other challenging high-value tasks. This can increase employee satisfaction and retention rate. Besides, since automation bots can work 7x24 continuously including weekends, this can increase company productivity & can provide a better experience to customers. Another key advantage of RPA is its non-intrusive nature, which the bot needs not to modify the coding in existing applications. Last but not least, it is a low-code tool. That means business users can become citizen developers to build automation by themselves.

AI/ML has also evolved to become a kind of off-the-shelf tool that a user without AI/ML knowledge can still make use of the technologies to build some simple cognitive automation. Especially during Covid to work-from-home, cognitive automation has been well widely adopted in many enterprises to maintain and even increase productivity. Examples like AWS Forecast (prediction engine) or Personalize (recommendation engine), they are packaged AI tools ready for ordinary users to use. With cognitive automation, useful data can be extracted from enormous amount of unstructured data which can be used to give meaningful recommendation or decision. In this way, a quicker action can be taken to the ever-changing economics.

Cognitive Automation combines artificial intelligence (AI) and process automation capabilities to improve business efficiency, conform to compliance and quality, and reduce turnaround and error rates. With Cognitive Automation, enterprises can automate decision-making and take quicker actions to ever-changing economics

Why did Cognitive Automation Fail?

Since the start of the Covid in late 2019, automation became the key to all CIOs. However, a major obstacle to automation is employee’s fear of being layoff. Because of this, those involved employees might create obstacles to the project or prolong the project duration.

Future of Cognitive Automation

Cognitive automation has evolved so that it can learn simply by observing an activity and emulating it without detailed instruction. Like adding machine learning (ML) to Automation to recognize new invoice hardcopy, there may be changes in layout format. With ML capability, Automation can self-learn to understand the new format and extract useful data. Automation project is not a purely technical project. It is about people, processes, and technology. It involves the transformation of employees. The Automation team should collaborate with the Enterprise Architecture team to identify the best-fit Automation projects. With the blooming of cloud adoption, automation pricing should be evolved to become AaaS (Automation as a Service) like SaaS so that a more flexible pricing model can be provided to enterprises to cater to the varying demand throughout the year.