Components of Intelligent Automation Platforms

By

As businesses continue to innovate and accelerate their digital transformation projects, countless software tools are being marketed to help achieve these goals.  These tools fill different needs, and many are being introduced as low-code or no-code options.  These terms can be misleading as they typically still require coding knowledge or technical backgrounds for more complex or enterprise use cases.  However, they do help companies deliver results quicker.  Intelligent Automation (IA) platforms are one of the enablement tools that drive greater efficiencies within organizations.  In this blog, we’ll review where to start and provide a deeper understanding of the components of an Intelligent Automation platform.

The most significant benefits to an organization tend to be reduced manual entry errors, increased speed to market, improved employee satisfaction, and increased capacity in the existing team to add more value to the organization.  Implementation of an IA platform positively impacts employee engagement and can lead to reduced costs and the ability to scale workloads quickly for company growth or seasonal work. 

So, where do you start?   

The UDig IA team will identify your automation candidates and work with you to formulate your business case. 

Intelligent AutomationAutomation Candidates

How do you pick the best candidates for automation?  At the simplest level, you should think of processes that your team does that are mundane, highly repetitive, rules-based, and essential business functions.  In addition to scouring your most visible processes, there are suites of process and task capture tools that are sophisticated and can provide additional ideas for automations (see Process and Task Mining below).   

Business Case & Return on Investment (ROI) 

When determining your business case, it is essential to evaluate things like the number of people who complete a task, the time taken by each person to complete a task from end-to-end, and the task’s frequency. These values will help you determine the ROI for your organization. Your ROI could include money or time saved, turnaround time improvements, full-time employee costs, opportunity costs, etc. 

Lets Dig In 

Now, lets touch on the components of the Intelligent Automation platforms which are a combination of many technologieused to automate business processes that are currently manual.  This overarching technology solution set can accelerate your digital transformation projects, assist in modernization and transformation efforts, and improve employee and customer satisfaction.  Intelligent Automation, sometimes known as Hyper-Automation, can include the following technologies: 

Process and Task Mining

There are multiple ways to capture a current state process.  One would be manual, where an analyst shadows the person (or people) as they perform the tasks that make up that process.   There are also process and task capture tools.  This can be something like a desktop recorder that captures every click and entry that the human performs, or it can be an agent that captures logs from the systems and combines that with the human entries to build a process map.   When the agentbased tools are used, more data will be captured over time to visualize any potential variations or exceptions in the process to have a very accurate view of the entire flow.  A human review of the process is still helpful when using some of the investigatory tools.   

Artificial Intelligence (AI) / Machine Learning (ML)

AI/ML models are typically used for a few components within Intelligent Automation.  The most prominent is intelligent document processing.  The ML models can extract information from documents without explicitly designating a location on a specific document for the data to be read.  The ML models require training by a human and can provide the level of confidence of the data that it reads being correct, then providing that feedback to the user. This training remains ongoing as more documents are read. 

Additional use cases for AI/ML are in the analytics and reporting functions management tools provide as part of the overall solution.  The analytics are instantiated in the initial analysis of the process to project potential time and effort savings and are carried through the measurement postimplementation to look for additional improvements.  In continuously running the process mining tools, AI/ML will find new process automation candidates. 

Computer Vision (CV)

Computer Vision is the computer’s ability to read screens in web pages, emulators, and remote desktop applications.  This technology utilizes Optical Character Recognition (OCR) to translate an image into text which can then be used to map data fields that need to be read, manipulated, or cut and pasted.  Depending on the software platform, there may also be AI associated with this technology to automatically determine data and data fields on the screen versus manually identifying the data locations on a screen. 

Optical Character Recognition (OCR) 

OCR technology has been around for many years. Still, it is the first step in converting a paper document into a format that can be used by a computer to be analyzed and categorized or used to be entered into another system.  This technology is typically used to take a scanned image or picture and convert it into text for further use and action.  

Natural Language Processing (NLP)

NLP is artificial intelligence that understands context and sentiment.  It can be related to extractions from a document, conversations in a chatbot, streaming information (i.e., social media), and devices or services such as Alexa, Siri, and Cortana.  This is a powerful addition to document intelligence because it allows intelligent reading of documents versus just extracting data from a field.  

Robotics Process Automation (RPA)

RPA is a software used to automate a process and become a new 24x7x365 digital worker, sometimes referred to as a bot. The tool is designed to mimic human tasks that include a definitive set of business rules that the bot can act upon (or be programmed to do). Many different companies sell the software.  

There are two types of bots,’ attended and unattended. Attended bots are launched by a user and can be run on a workstation, server, or cloud service and typically used when human intervention is necessaryOn the other hand, unattended bots run on a server or in a cloud service and do not need to be monitored or have any human intervention. 

App / Business Process Management System (BPMS)

There may be a need to automate a longrunning business process due to complexity and workflows that may include many human/robot interactions.  Creating this complex process flow may require the use of an app or BPMS software solutionRPA and IA can be combined with BPMS systems to provide automation value to new or existing workflows.  

Document Intelligence

There are multiple different ways to extract data from a document.  In a structured document, identifying the fields and telling the software where to look for the data you need is straightforward.  The assumption is that this document is static, and the computer will always find the data in the same place on every document (i.e., standard forms). 

You will need to use more advanced methods of extracting data in a document containing semi-structured or unstructured data.  These methods include AI, ML, and NLP to enable the computer to read and understand the document.  If the document is handwritten or has not yet been input into a digital system with character recognition, you will also need an Optical Character Recognition (OCR) engine to assist.    

Automation Insights (pre and post)

The first challenge is choosing a process automation candidate that is a good fitWhat is the expected return on investment from hours savedWhat is the opportunity cost of not being able to complete more worthwhile projects or work efforts?  Will this help with employee turnover or improvthe customer experience The capabilities that come with the management utilities assist in the pre-automation and discovery of the most valuable automation candidates and monitor them post-implementation for continuous reporting on effectiveness, time, and financial savings. Analytics models are used to identify candidates and make recommendations for enterprises to make the best of their IA platforms. 

This is a point-in-time view – more functions and capabilities are continuously being developed and integrated into the platforms. 

At UDig, we evaluate your business process to ensure that we understand the current state process and make recommendations on any efficiencies that can be gained, even before utilizing automation tools.  After that initial review, if a process is deemed a good candidate for automation, we will architect a solution using the modules mentioned above to provide you with the most stable and resilient software implementation.  From there, working with your team, we will develop the software to deliver the highest value for your use case.  Please feel free to contact us to review your potential use cases and see if an Intelligent Automation solution is right for your company.   

About The Author

Chris Lara is the Senior Director of Intelligent Automation. He is a 30-year technology industry veteran working with clients of all sizes with a local to global reach. Chris is passionate about utilizing technology to drive value to all business functions and employee and customer experience.