RPA & AI (UiPath and Machine Learning)

Source: Deep Learning on Medium

What is RPA (UiPath) ?

UiPath logo

Robotics Process Automation (RPA) is a technology that helps you in automating your business process. It is a technology which mimic human actions in interacting with digital system. Hence, RPA making human’s less “robot” by automating tedious, repetitive, time-consuming and non-value added task.

To understand what‘s RPA in a more intuitive way :

What is RPA and what isn’t RPA

What can RPA do for you?

What RPA can do

UiPath is one of the top leaders which develops Robotics Process Automation platforms. The reason that this blog is written based on UiPath Platform is as follows:

  1. Easy to learn (various of available tutorial)
  2. User friendly (drag & drop coding interface)
  3. Great community and training support (UiPath Forum & UiPath Academy)

Here’s a demo video of UiPath Platform.

By deploying RPA into your business process, it helps your business in reducing operation cost , eliminate human error and saving time. Furthermore, it allows your business model to growth by freeing up skilled resources for strategic task and create new jobs.

Another benefit to consider of deploying RPA into your business process is you don’t have to invest heavily to automate essential processes. Without any additional installations, RPA still integrates well with your company existing IT infrastructure. In simple term, RPA is easy to implement and scale.

Sample RPA workflow

What is Artificial Intelligence (AI) ?

Artificial Intelligence is one of the most cutting edge technologies which makes machines react with human-like intelligence. AI term coined by a researcher named John McCarthy in 1956 when he hosted a workshop called Dartmouth Summer Research Project. John McCarthy defined AI as “ The Science and Engineering of Making Intelligent Machines“. It is important to understand that AI is an umbrella term that involve various sub-fields in computer science. As of today, AI has being focused on the sub-field of computer science namely machine learning (ML).

Markets nowadays invest on AI deploy AI for making digital system having the capability of “human reasoning”. Subsequently, they uses this capability to provide better services and products.

Machine learning can be understand as in a way to implement artificial intelligence. You can think of machine learning as a subject in computer science where you learn about the design of algorithms. Hence, for people who already know about machine learning, you may be already familiar with those algorithms such as decision tree, support vector machine, K-means clustering and etc. These algorithm learn and interpret data to allow machines to perform certain cognitive tasks.

Decision tree classifier

Although conventional machine learning algorithms performs well, most of these are still very machine like. This means that these required lot of domain expertise, human intervention in order to maintain it performance. For example, when a machine learning algorithm returns a bad prediction result, it require a on field machine learning engineer to step in for adjustment.

Hence, there is another powerful sub-field of machine learning algorithm is been introduce to tackle real-world complex problem. Deep learning (DL), a sub-field of machine learning algorithm which inspired by hows human brain works. Deep learning algorithms are modeled after the human brain, designed to recognize patterns.

Artificial neural network

Why Deep Learning?

The rapid growth in technology especially in internet sector has generated massive amount of data. Conventionally machine learning algorithms have limitations in training well with enormous amount and deeper dimension dataset. In contrast, deep learning algorithm works tremendously well with more data .

Why deep learning now?

Deep learning algorithms come along with automatic feature extraction capability. Lets take an example here, when performing image classification task, convolution neural network layers learn to perform feature extraction itself based on the loss function calculation. Throughout a number of iteration, the convolution neural network layers will gets better and better in extracting image features.

What is RPA & AI?

As for now, you probably have understand what is RPA technology and AI technology and you are interested to know how these two technology can work together in order to improve your process, service, product and etc.

Making your software bots intelligent

RPA & AI are two different technologies which distinct in solving separate sets of problems. For instance, RPA is a process-driven technology to automate repetitive, rule-based and non-value added task in business processes. In contrast, AI is a Data-Driven technologies where it learn from previous computational outputs (data) to make reliable decisions and results.

Both technologies have different strengths in solving real-world problems. As mentioned for RPA, it aims to automated repetitive, time-consuming business process and hence, RPA is more on “execution side”. Vice versa, AI aims to enable machines(computers) to perform intellectual tasks as problem solving, decision making, perception and understanding human languages(Natural Language Processing). In short, AI is more on “thinking” side.

RPA and AI objective

Combining them together!

Think of this way, RPA is capable in execution and most of the time RPA is just “doing” by the instruction explicitly provided. RPA lack of the capability to “think”. In other way round, AI is only able to perform intellectual tasks such as decision making and etc. AI technologies itself unable to perform “execution” without integrating into another technologies.

Think of RPA is your “hand” and AI is your “brain”, by integrating AI cognitive capability into RPA technologies, you own a virtual employee to execute your business processes.

Virtual employee

As for my thought, combining RPA & AI technologies are two ways improvement strategy. AI technology enhances RPA performance by augmenting cognitive capability into it. AI technology allows RPA to further automated business processes and mitigates the needs of human intervention in between huge processes.

Take an example here, resume screening process in HR department. Without AI technology involves, the screening process will stop after right after it downloaded all the candidate resume and saved in some folder. After this, experienced HR employees have to come in to screen through all candidates resume according to the job requirement for different position. Imagine how tiring only for the screening part if there is 10 different job opening and each of them have 100 candidate submitted their resume! So, to further improve this resume screening process, we can integrate with AI technology. Training a deep learning model that learned from past experience(resume dataset) to classified on each incoming resume whether is a fit or not. The whole process can be fully automate start from collecting all resumes until scheduling for first interview session.

RPA & AI example : Resume filtering

RPA & AI does not only apply on leveraging AI technology to enhance RPA technology. RPA is good at automating tedious, repetitive and tiring process. When it comes to the training of machine learning models, the fundamental and most important ingredient is data. So, how can we get data? well, probably you will tell me you already have enormous amount of data. My question here, is the data structured? Is there any processing needed to be done?

RPA can play an important role here to automate the process of data pre-processing and even the automation of data collection process! Take another quick example here, assuming you wanted to train a recommendation engine machine learning model. That model is aim to recommend tourist where is the best place to visit in South East Asia and you will need to scrap data from multiple travel website such as TripAdvisor, Expedia, Booking.com and etc. So you will have to go to each website and scrap all the data!? Nope, you can use RPA here. RPA allow you to automate the whole process of data collection in a structured form (.csv, .xlsx and etc.).

Summary of this blog post:

With the integration of RPA & AI technologies into the business process, the benefits that come with it are:

  1. Greater cost reduction
  2. Able to automate higher-order processes
  3. Better efficiency and accuracy