Infosys’ blog on industry solutions, trends, business process transformation and global implementation in Oracle.

« June 2018 | Main | August 2018 »

July 14, 2018

Robotic Process Automation - Capabilities Overview

 

Robotic Process Automation - Capabilities Overview

Understanding the Basics

 


 

 

Introduction

Every few years, IT / ITeS industry is seeing a new product or technology which bring exciting new UI features, capabilities for users to configure the application easily and a lot of new buzz words and concepts for technology enthusiasts to get accustomed of.

Robotic Process Automation and Artificial Intelligence are 2 such buzz words which have excited fast growing organizations and IT industry equally and are really catching up fast.

While for the organizations, it opens up new avenues to achieve higher operational efficiencies and cost reduction, eventually impacting the bottom line; for IT industry it opens up new horizons to increase the client base with new offerings and increase the technological footprint.

In this blog, we'll focus on the overview of Robotic Process Automation - the basic understanding of RPA, the types of RPA and what encourages the fast growing organizations to go for it.

 

What is RPA?

In this competitive consumer market, organizations face a perpetual challenge of moving swiftly while keeping the costs of operations low while increasing the consumer satisfaction level and service offering quality.

Though organizations are aware that to reduce the costs, they have to achieve higher operational efficiencies; however, there is a direct impact on bottom line if they hire more staff to achieve that. Increasing the working hours, paying overtimes to existing staff again dents the profits. To alleviate these challenges, organizations are finding their savior in 'Robots'.

In simple words, RPA tools (Robots) emulate the manual steps as done by users across and through applications from UI entries. RPA tools operate by mapping a rule-based workflow process which the "robot" can follow to the 'T'.

An important point to note here is that these Robots can be implemented agnostic of system or application. These Robots can be as simple as batch file upload automation to as advanced as a cognitive automation which has self-learning, variable format processing capabilities. Processes can be triggered manually or automatically to:

·         Populate data across different systems and modules within them

·         Run queries on a scheduled basis and perform data reconciliation

·         Generate and distribute reports

·         Auditing of large volumes of data

·         Trigger downstream activities and processes

Per proven studies conducted by various institutions and agencies, it has been identified that in large organizations, there is typically a scope of saving 20-40% of workload for employees with help of automation and imagine what levels organizations can achieve with employees having 20-40% more of their time to focus on value added tasks.

 

Types of Automation

As mentioned above, there is plethora of tasks which Robots can do starting from simple, mundane activities like data entry to super complex activities like generating a dynamic response to user query based on machine learning and cognitive abilities.

There are different stages or levels of Intelligent Automation:-

·         Digital Worker - As evident by the name, this is the primary or entry level of automation an organization can move ahead with and still achieve efficiency gains.  This is the typical Robotics Process Automation tool which can perform tasks like:

o   Data entry

o   Running functions in excel towards data validation

o   Triggering customized emails with preset content or standard templates

o   Data comparison

o   Setting up reminders

o   Batch processing and populating mapped fields

o   Queuing and Assignment

 

·         Digital Reader - This is a secondary level of automation alternatively referred as 'Cognitive Automation'. Robots at this stage can perform tasks involving:

o   Machine Learning

o   Pattern or Keyword based recognition which is evolving over time as Robot sees and identifies more patterns / keywords

o   Data processing across variable formats

o   Dynamic queue assignment based on patterns

o   Complex analysis based on continual learning 

 

·         Digital Talker - This is an automation offering which focuses on providing a more interactive experience. Alexa from Amazon, Google Home are very popular examples of this. These robots are also called 'ChatBots' as they perform somewhat similar tasks as the previous classification of Digital Reader'; however have additional text and voice capabilities and are more communication focused. Robots at this stage can perform additional tasks involving:

o   Predictive Analysis

o   Customer Servicing

o   Query Resolution based on pattern or keyword based recognition which is evolving over time as Robot sees and identifies more patterns / keywords

 

·         Digital Thinker - This is the advanced level of automation which is the classic Artificial Intelligence. Artificial Intelligence tools are somewhat comparable to humans in terms of intelligence and have their own IQ. Currently, the IQ of these AI tools is significantly lesser than that of humans. Per studies performed in 2016, IQ of Google's A.I. (47.28) is nearly two times that of Siri (23.94), however a six-year-old child beats both of them when it comes to smartness and thinking capability. An average person's IQ is in range of 85-114.

As the IQ of these applications or tools increase, to a certain point it'll be beneficial for people and once the IQ surpasses that of average human, then we all know what will happen - we all have seen sci-fi moviesJ.

Nonetheless, the Digital thinker can perform below activities in addition to the activities listed for previous categories:

o   Predictive Analysis based on cognitive learning and complex algorithm 

o   Complex Mathematical Analysis

 

 

RPA Benefits

 

RPA_Diagram_1.jpg

Conclusion

Organizations need to be smart enough to understand their IT Landscape, business process steps and identify the correct tasks which can be automated with help of Robots. Although other competitors might be at a higher level of automation, an organization needs to be realistic in its approach and move with Automation stages through a proper strategy and careful planning to reap the benefits of automation.

 

Reference

1.       https://www.cnbc.com/2017/10/02/google-ai-has-almost-twice-the-iq-of-siri-says-study.html

2.       https://en.wikipedia.org/wiki/Robotic_process_automation

July 5, 2018

GDPR - Its impact on CX


General Data Protection Regulation (GDPR) came into force in European Union (EU) member states from 25th May 2018. It has far reaching ramifications for businesses and organizations given that data is ubiquitous and all businesses today rely on customer data to remain competitive in their industry and relevant to their customers.


In this blog we will examine some of the challenges that businesses in certain industries can face and what businesses can do about it. GDPR restricts itself to personal data thereby limiting its regulatory reach to all such companies and organizations that are serving direct consumers of their services.


In this era where advertisements on social media, advertisements on web pages, advertisements on mobile applications are personalized by gathering and processing information about the specific user how can companies that use these media of connecting with their customers continue to send pertinent communication/messages to their customers.


In retail ecommerce customers are shown recommended products using association rules and recommender systems. This is possible because the company keeps track of customers past purchases (past buying behaviour) so as to recommend new products to the buyer.


After implementation of GDPR the following can happen


  • The buyer can refuse the ecommerce company to control and process his/her data. This at once nullifies all the investment it has made in processing this buyer's data as it is brick-walled from its customer.
  • On the flip side it gives a level playing field to other ecommerce companies as every buyer out in the market is anybody's customer. In short, customer loyalty will be short lived.

So how can organizations and companies insulate themselves from losing out their customers? The answer is simple and has stood the test of time - roll out the best service to each customer whether the customer is buying from them for the first time or the hundredth time. Companies will have to relentlessly satisfy customers in every transaction so that customers willingly share their data. Period.

 According to Epsilon research, 80% of customers are more likely to do business with a company if the company provides personalized service. With a possible destruction of customer data after completion of transaction as stipulated in GDPR


  • It is difficult for companies to personalize their offerings to "customers".
  • Customer profitability KPIs like Life time Value(LTV) may not be meaningful anymore as the same buyer is a new customer each time if the buyer chooses to annul his/her personal data after completion of every transaction.
  • Newer catch-phrases like Customer Journey Mapping fall off the grid as the "traveler" in the "journey" is temporary and companies may not even know the "traveler" i.e. the customer.

So how can companies personalize their services to customers? Prudent companies can anonymize customer data by encrypting it immediately after sourcing it. Though this will not help them decrypt to find the specific customer the still company has some sort of a handle on its customer.

Companies in the financial services rely on accurate, updated and complete customer data to discern genuine customers from fraudulent ones. To keep good customers separate from bad ones companies will have to be innovative to "pseudonymize" customer data.

So how does this work?

GDPR only regulates personal data and not transactional data. So financial service organizations will have to "pseudonymize" customer data using new technology mechanisms (which may or may not exist today) so that customer data is also treated as transactional data. Such transactional data can then be trained using Machine learning/Deep learning algorithms to spot fraudulent customers from reentering the financial services market.

All data is stored in servers and server farms on the cloud or in in-house data centres. As the financial cost of misdemeanor in following the GDPR is very high (ban on customer data processing and a fine of up to higher of €20 million or 4% of the business's total annual worldwide turnover) the IT and ITES industry may also not be immune to impacts. The following impacts may be notices


  • There may be instances where the processor of the data (the organization that defines the how and why of customer data) may move the data on-premise thereby playing the role of controller of data as well. The controller is the one holding the data like AWS.
  • Small businesses may be tempted to move from cloud to on-premise to reduce chances of data theft or rework their contracts with data controllers to insure themselves.

With the widespread use of data science and machine learning in business, companies would have to be very diligent in deleting customer data from training data that is used to build supervised algorithms if a customer asks for deleting his/her personal data that is part of training data. If many customers follow suit then the model so built is itself now rendered inefficient as the training data has changed and patterns have to be learnt again. Companies will have to keep their learning algorithms and models updated regularly so that their outputs are pertinent.

GDPR puts the onus of processing data on companies and organizations and awards private individuals complete rights over the way their data can be stored and processed. As individuals become custodians of their data they may choose with whom and for how long they may share their data. Is it possible in the future that large groups of users form cartels and charge businesses for using their data?


Subscribe to this blog's feed

Follow us on

Blogger Profiles

Infosys on Twitter