Cognitive Automation – Intelligent Automation – Part 2

Cognitive Automation – Intelligent Automation – Part 2

As mentioned in the previous article, Intelligent Automation is the intersection of process automation, and artificial intelligence (AI) technologies. In part 1, we talked about the role of RPA as a driver in improved quality, improved efficiency, and lowering the cost of business.  This is achieved by software that can script every day human-computer tasks which can run faster than several humans, and without data entry errors. In this article, we discuss how Artificial Intelligence furthers this important digital transformation capability.  However, Artificial Intelligence is a VERY broad conceptual term. I have yet to find one consistent definition, so we’ll take our hand at a definition.

Simply defined, Artificial Intelligence (AI) is the theory and development of software that can “think” and act like humans, i.e. mimic cognitive functions. In the context of today’s capabilities, to “think” is interpreted as software capabilities built on mathematical and probability models to predict values, recognize patterns, and identify relationships between data.  There are two subfields today that apply AI research into practical solutions: Machine Learning (ML), and Deep Learning (DL).

Furthermore, Machine Learning is a specific, applied subfield of AI that focuses on “learning” from data without specific programming.  Deep Learning is a subfield of ML that is focused on learning data representations, i.e. pattern recognition, as opposed to task specific algorithms. These techniques use historical data to “learn” to make a model. Then these models are used to make decisions with new data.  These models can continually learn by regularly updating them with newer data, thus evolve. NOTE, this is an important feature in maintaining AI models, and a future topic on an AI Operational Architecture (AIOps).

Sounds great, but why do we care?  AI is part of the core capability for a data driven organization, i.e. using data to drive the business.  How often does your business rely on individuals to review data, textual reports, and dashboards to make decisions in the business? Some examples include:

  • Predicting close rates on future sales in the pipeline.
  • What is the proper answer to a specific customer service question?
  • Is a particular transaction fraudulent or not?
  • What products should I recommend to a customer online?
  • From our repository of thousands of documents, which ones discusses a specific concept?
    (note, this isn’t a text search).
  • In a manufacturing line, which components are going to fail in the next 3 weeks?
  • How much product should I order in the next 30 days?

Humans are inherently limited on how much data they can retain and process mentally. We are also influenced by how they learn, i.e. who/how/where you were taught. We typically learn in business on-the-job based on general rules or high-level understanding of data by a colleague, but our decisions don’t necessarily reflect the actual data. ML/DL models do not have this restriction. These models process huge amounts of data and can repeatably make predictions.  The goal is to meet or exceed humans in the decision process and perform tasks faster.

Each one of the examples above have several things in common.

  • First, they are business-oriented problems that are addressed every day.
  • Second, they all have underlying data, typically a lot of data, often originating from multiple data sources.
  • Third, recognizing these and other similar business processes is the first step in defining an AI solution, and building a strategy.

One of the key steps in an Intelligent Automation Strategy is starting with recognizing the business opportunities for automation.  The company’s strategy is a good starting point. The other approach is identifying business processes supported by data as candidates for automation. If you’re new at it, it can be challenging.

Many companies believe in the value of machine learning, but often struggle to use it. This could be a gap in understanding, lack of skills, or lack of strategy. In other cases, projects get stuck in pilots. Votum and its partners provide consulting services to help companies understand and guide companies in their Intelligent Automation journey. It all starts with a conversation.

Contact Votum to understand how we can help in your Intelligent Automation Journey.


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.