Robotic Process Automation

From New Media Business Blog

Jump to: navigation, search




Robotic Process Automation is software that mimics human behaviour for business process delivery. It does so by interacting with existing technology at the user interface level, making it significantly simpler than typical machine communication through Application Program Interfaces (API's). [2] This capability allows RPA to have greater speed and predictability, with less process disruption when compared to traditional process transformation as seen on the graph on the right.

RPA can reduce the number of employees that are doing repetitive, high-volume, rule-based tasks, and contribute to making jobs more interesting and rewarding. The most significant benefit of RPA is the time savings and consistency, reducing process time by an average of 76%. [3]

Emerging Growth in the RPA Market

The graph below shows the RPA Market in North America by Application


Robotics used to be big, immobile machines doing very repetitive tasks. With other emerging technologies such as artificial intelligence and machine learning, the market is expected to grow 20 times by 2024 to a $4 billion industry in North America alone. Some of the industries with the biggest potential include banking, financial services, insurance, telecommunications, IT, manufacturing, retail, and healthcare.

Advantages of RPA

The Rise of Machines - Why Automation is Different This Time[5]

While automation is not new, the new wave of automation has significantly greater capabilities and implementation opportunities. :[1]

Cost reduction: The biggest advantage of RPA is the reduction in human capital spending. Robots are available 24/7 and have no salary, turnover, or benefit costs. This increased availability can also assist in areas with staff shortages.

Consistency and Auditability: [2] RPA also provides a greater level of quality control since every step of the process is programmed into the robots. They will consistently follow configured algorithms and the programmed business logic. As a result, there is an increased level of operational and business predictability. This is advantageous in roles that can be negatively affected with bias or variation and have clear, rule-based decision flows.

Compatibility with Existing Systems: RPA mimics human actions, interacting with existing technology at the human interface level, which means it has a non-invasive nature. The implementation of RPA into existing processes can be accomplished fairly quickly compared to most technology, assuming that the processes are mapped out clearly. Furthermore, it does not require programming or a vast amount of technical resources.

Efficiency: RPA provides efficiency and can increase speed to market. Furthermore, these efficiency advantages can also be seen on the front lines by improving customer service. It can significantly reduce response times to customer requests or complaints.

Re-shoring: Among increased adoption of RPA, there is a new trend in labour practices. As opposed to offshoring, many off-shored jobs are now returning back to the home country, with companies taking advantage of robotics, which is consistently cheaper than even off-shored labour. Although this trend may only be adding few additional jobs for people to oversee and manage these new facilities, this can reduce the environmental impact that manufacturing products overseas has.

Limitations and Risks of RPA

Systemic Issues: [1] While the consistency of RPA is a strong advantage, it can create systemic-wide issues if a certain step or rule is missed or not programmed adequately. While human error may be more likely, it is unlikely to be widespread. Nonetheless, this can be mitigated through human intervention or using another layer of RPA to check that outputs are being processed correctly.

Poor Processes; Poor Results: There is also the potential risk of RPA being used as a tactical fix for inefficiencies or ineffectiveness within the business. Due to the relatively low cost and simplicity compared to many other new technologies, it may become a chosen solution without a full root-cause analysis of underlying problems.

Literal, Rule Based Tasks: [1] Can a robot really replace human creativity and flexibility? The answer is no. RPA is limited to repetitive, rule-based tasks that must be handled in a consistent and black or white manner. RPA can only make decisions based on clear data and criteria that it has been programmed to analyze. This also increases the risk of missing unwritten rules in the process that aren’t necessary due to human intuition.Nonetheless, RPA combined with thinking or learning capabilities is becoming more common with advancements in artificial intelligence and machine learning.

Inflexible: Due to the necessary programming of all criteria and rules, the software may fail if business processes change. To mitigate this risk, changes must be planned, communicated, and tested using strong IT governance structures in advance of the change to accommodate time for any changes in the robot programming. Nonetheless, newer solutions have increased capability to accommodate simpler changes without issues.

Security, Privacy, and Governance: Before RPA can be implemented, involvement from corporate compliance, risk, internal and external audit is recommended to ensure adequate risk assessments and governance frameworks are in place. Many countries have also taken steps to ensure privacy compliance is built into their RPA framework. This includes the security risks involved with the potential to hack into robots.

Limits Social Interaction: The largest limitation with robotic technology is perhaps in the service industries. While technologies are in place for robots to act as your barista or server, can robots replace the human need for social interaction?

While most people would agree that the replacement of many bank tellers with ATMs was beneficial, how far can this replacement of people with objects go? Often, this depends of the type of experience the consumer is seeking. Services that are focused on convenience and efficiency, such as a fast food restaurant, may be able to benefit from implementing RPA, while those services that are highly differentiated or high-quality, such as a fine dining restaurant, may not be able to utilize the same technology. Furthermore, it is important to note the ability to have robots work in conjunction with humans. The robots can do the repetitive work, such as flipping burgers, while interactions with the customers or more creative work, such as plating, can still be done by humans.

Effect on Canadian Labour Force

High Risk of Automation

According to the Talented Mr Robot report, 42% of jobs are at a high risk of automation by existing technologies. These jobs require less education and have lower salaries. Occupations at high risk include cashiers, salespeople, truck drivers, food counter attendees, and administrative assistants. [3]

Low Risk of Automation

That being said, 36% of jobs are at a low risk of being replaced by automation. These jobs have an average salary of $61,000 per year, compared to the $33,000 per year average salary earned by those jobs likely to be replaced by automation. Jobs that require a high degree of flexibility, creativity, adaptability, intuition, or relationship and network building are at a lower risk. This includes middle to upper management roles, nurses, and teaching roles. [3]


Analysis of Canadian Occupations Affected

The graph to the right shows that most jobs are at a low or high risk for being automated. This indicates that RPA is still limited to repetitive and rule-based tasks, and confirms that there are numerous limitations that make several jobs safe from automation risk. Furthermore, it indicates that jobs that will be automated tend to have lower incomes, which presents an opportunity to improve the standard of living by re-educating the workforce for better jobs. However, the risk of simply increasing the unemployment rate is high. Employees at higher risk are significantly younger and less educated.

Furthermore, it is important to note that of the almost 1.5 million new jobs emerging between 2014-2024, 70% have low-medium risk of automation. [3] This validates that increasing automation is resulting in emerging jobs. These emerging jobs have higher incomes and also require more education.

To reduce the risk of an increase in unemployment, it is necessary for Canada to uncover and capitalize on technological strengths. Re-training the workforce for emerging work can ensure that entrepreneurs have access to the talent needed in Canada. This can create a more innovative economy and continue to grow Canada's technology industry's reputation.

RPA Impacts on the World

As of right now, many corporations rely on the labor force of developing countries. This has been going on for decades with large corporations such as Nike utilizing a low cost labor force to increase their profit margins. Aside from the labor force of manufacturing, many companies now outsource their departments in other countries to drive down cost as well.

With the implementation of RPA in many companies, this is a prime opportunity to reshore jobs to increase control over certain functions or improve quality, while receiving the same cost benefits. The increasing trend towards reshoring has a positive impact on jobs with poor working conditions as technology will eliminate the need for cheap labor in developed countries. But a necessary question to consider is, “What will happen to developing countries and their labor force?” It’s no secret that the future of our world is filled with technology, and many countries will need to begin adapting. Not only will developing countries be pushed to catch up and educate their young, developed countries will also need to push kids to develop new skills in preparation to step into high-skill occupations that have yet been automated. [4]

Potential Issues of Widespread Automation

Rise in unemployment

One of the biggest fears with RPA is the threat of increased unemployment. Currently, RPA is targeted towards replacing jobs that require repetitive tasks and with many entry level jobs performing similarly, it is becoming an increasing concern. While RPA may not take over these jobs overnight, there is still the issues of a gradual job loss in these sectors. [5]

Reduction in spending power

With increased unemployment, individuals will have less money to spend and will have decreased financial stability. People will not have the money to spend on what they need and there will be an decreasing amount of money available in the economy. [5]

Increase in inflation and recession

Following a reduction in spending power, the last step of this process would be an increase in inflation and an increased risk of a recession occurring. A more extreme, worst-case scenario of widespread automation could be an economic collapse.[5]

Challenges of Implementation

Employee Resistance and Onboarding

With any changes in a company, there is a big role of managing employee expectations. Implementing new technology can be stressful for employees as they might experience shifts in their responsibilities. It is essential for company leaders and executive sponsors to frequently communicate expectations through the implementation process to lead to a successful adoption. In addition, fostering a culture of innovation within the company will only further accelerate this adoption.[6]

Choosing the right processes

The automation capabilities provided by RPA are ideal for tasks that are repetitive, rule-based, high volume, and do not require human judgement. This can include activities such as data migration and copy-paste tasks. RPA implementation is especially difficult though with business processes that are non-standardized and require frequent human intervention in order to execute. For some businesses, processes might be written down and be “official processes”, but in reality, employees may be using a more heuristic or practical approach to the problem at hand. In those cases, it may be difficult for the business to utilize RPA. Typically, these more complex tasks include interacting with customers and developing human relationships. While it is an upfront time investment, it’s important for companies to determine which of their processes are suitable for RPA so that automation runs smoothly.[6]

Setting Realistic Expectations

This is arguably one of the biggest obstacles when it comes to implementing a new technology such as RPA. Instead of seeing RPA as the panacea for operational problems and broken processes, organizations need to recognize the limits of what RPA can and cannot do. Decisions regarding the technology need to be made on an individualized and company-specific basis, as RPA’s functionality, implementation timeline, and operational results will vary between different companies. Maintaining company-wide discussions about expected results will allow organizations to make the most of RPA and its benefits.[6]

Ongoing Debates and Other Issues

RPA is a new technology that many people have a limited knowledge about, even though at the end user’s level they may require a certain level of technical knowledge. There is a lack of knowledge from the user’s perspective for the tool making it difficult for users to adapt. With a limited understanding, it is easy for people to adopt a culture of fear regarding RPA, considering it as a destructive technology rather than a disruptive, support tool that it is currently marketed as.

Considerations of Implementation

Organizations will have many questions to answer when implementing RPA. A significant question they will need to answer is what processes are they choosing to implement RPA. Process changes in current organizations already take significant time, cost and effort. Now consider implementing a new technology, how much more time, cost and effort is needed for a company is a huge question to answer as well on top of organization structure changes. Perhaps more importantly a concern employers will have to settle when considering implementing RPA is the fear of job loss for employees. The amount of change is already unknown and daunting to companies, and even more so from the perspective of the employee. Employers will have to plan for jobs that employees can transition into in order to remove some of the fear regarding job loss. More importantly, industries looking to take advantage of this technology will have to start investing and creating roles that employees can grow into to successfully adopt RPA.

Adoption of Robotics Around the World

Robot density per 10,000 employees in manufacturing (2015)

The Republic of Korea, Singapore, and Japan currently hold the top spots in terms of robot density (number of robots per 10,000 employees) in the manufacturing industry. They respectively have robot densities of 531, 398, and 305, with the world average being only 69[7].

The automotive industry is responsible for the majority of industrial robot shipments[8]. However, robot sales to the electrical/electronics industry (including medical equipment) have been significantly growing, with sales in 2016 increasing by 41%—higher than any other industry[9].

Global annual shipments of industrial robots are set to see a 16% growth rate through 2025[8]. According to the International Federation of Robotics, Japan is currently the world's predominant industrial robot manufacturer, supplying 52% of the global supply[10].

Current Applications of Robotic Process Automation

Automotive Manufacturing

Innovative human-robot cooperation in BMW Group Production


German automaker BMW has been using robots to help automate processes involved in vehicle manufacturing. In doing so, they have increased the efficiency of their workforce in a collaborative manner. Automation allows workers to apply cognitive skills that robots do not have, while the robots handle the ergonomically unfavourable and strenuous tasks[11].

Tesla Motors

Unlike BMW, Tesla Motors has had a more aggressive adoption strategy towards robotic automation. They have been been ramping up their usage of robots during vehicle production under increasing pressures to meet purchase demand. Their Model 3 production process is vastly more automated than previous models, with processes such as welding (which is typically done by humans) now being completely automated[12].

Tesla wants to move towards an inhuman future. One of Elon Musk’s goals is to radically transform not just auto manufacturing, but all manufacturing, by using a higher degree of automation at speeds that are beyond human capability. In fact, humans would not be allowed in the production line itself as the robots’ speeds are too dangerous to be around[13]. Humans would only oversee the robots and ensure everything is running at peak efficiency[14].

Tesla Model S Manufacturing [15]

Autonomous Delivery Robots

Starship Technologies' Delivery Robot [1]
Startups such as Starship Technologies[1], Dispatch[2], and Marble[3] have designed self-driving delivery robots that deliver small loads within small radiuses on sidewalks. The robots move at pedestrian speed and can navigate around objects and people in busy urban environments[4].

Delivery startups such as Postmates and Door Dash are using the robots to deliver parcels, groceries, and food from restaurants. Customers simply make their requests via a mobile app, and when the robot arrives, customers can unlock the cargo bay with a code.[5]

The aim of these delivery robots is to make local delivery faster, smarter and more cost-efficient[1]. It is estimated that human delivery costs account for up to 80% of expenses occurred by companies like Instacart and Postmates, so using these robots can result in tremendous efficiencies[6].

Health & Medicine

In healthcare, IBM Watson is being used to assist physicians in identifying treatment options for their patients. Physicians are able to describe symptoms, allow Watson to form and test hypotheses, and receive a list of recommendations along with the confidence level for success for each option[7]. Watson can examine millions of treatment guidelines, electronic medical record data, research materials, clinical studies, journal articles, and patient information[8] exponentially faster than any human. (Scientific research grows at a rate of about 8,000 new academic papers per day—it is virtually impossible for doctors to keep up with findings.[9]) With the found data, it can then find meaning within it and make decisions that are evidence-based and free of biases or overconfidence.

In an analysis of over 1,000 patients, Watson found the same treatment options that doctors recommended in 99% of cases. Furthermore, in 30% of cases, Watson found new treatment options that doctors did not find[10].


ROSS Intelligence

ROSS Intelligence[11] is a platform based on IBM's Watson that helps legal teams sort through case law to find details relevant to new cases. This process normally takes days to weeks, but lawyers are now able to ask ROSS questions in natural language which then sifts through over a billion documents per second. Firms are able to find 30% more relevant information in less time, freeing up time and reducing costs.[12]

Furthermore, ROSS learns from feedback and gets smarter over time; in other words, it is learning to understand law and could perhaps one day do more than legal research.[13]

Logistics & Warehousing

Amazon's Kiva Robots [14]
In 2012, Amazon purchased Kiva Systems for their automatic storage and retrieval robots[1]. Warehouse employees used to walk around the warehouse to find items on shelves. Now, the Kiva robots bring the shelves to the employees who simply stand in one place. Because there is no more need for walkways, shelves can be packed tighter, increasing total storage capacity by 50%[2]. Order fulfillment time takes just 15 minutes compared to the previous 60-75 minutes, and warehouse operating costs have been reduced by 20%[3].

Although Amazon ceased to renew contracts with other companies who were using Kiva robots before the acquisition[4], several other alternatives have been developed by companies such as Locus Robotics[5], Geek+[6], 6 River Systems[7], and GreyOrange[8].

Future Applications of Robotic Process Automation

Autonomous Surgery

While there have been robots such as Da Vinci have been able to perform surgery on humans for years, they have all required a human operator. Researchers from the Children’s National Health System in Washington, D.C., and Johns Hopkins University have successfully performed autonomous operations on pigs using STAR, or “Smart Tissue Autonomous Robot.” No previous robot has performed soft tissue surgery until now. [9]

STAR uses a precise Kuka robot arm in combination with near-infrared and light field (plenoptic) 3D cameras and is able to independently choose the best way to accomplish a task. However, surgeons still need to mark key portions of the surgical areas so that STAR can see them.[10]


SAM [11]

SAM (Semi-Automated Mason)

SAM (Semi-Automated Mason) is a brick-laying robot that is able to place approximately 100,000 bricks per month. A job that would normally take over half a year for a construction worker to do can now be completed in just one month. However, that’s not to say that construction workers will be replaced. SAM is intended to work alongside humans to install bricks, making the job less backbreaking while increasing productivity. Construction workers continue to use their knowledge and skills while the robot handles the repetitive physical labour.[1]

Disaster Relief

Honda has developed an experimental humanoid robot for disaster response called E2-DR. E2-DR is strong, nimble, and can get wet without exploding. Its intended usage is to be a first responder to disasters in infrastructures such as nuclear plants, and will be able to[2]:

  • climb stairs, stepladders, and vertical ladders
  • see in almost any lighting conditions
  • move in narrow spaces
  • climb over pipes
  • pass through closed doors
  • move over scattered debris
  • perceive the environment
  • prevent catastrophic falls if it loses power during movement
  • handle real disasters (dustproof, splashproof, and can work in extreme temperatures)[3]

Electronics Recycling

Meet Liam [4]


Apple developed a recycling robot named Liam to disassemble iPhones and recycle parts. Traditional recycling methods involve shredding devices and using magnets to separate some parts. However, the result is that you often get different scrap materials mixed in. Liam separates the components of an iPhone with robotic precision one part at a time such that every piece is sorted during disassembly. This results in a recycling method that is easier and takes just 11 seconds per phone.[1]

As of now, Apple is unaware of any other company that disassembles electronic products in a similar way that makes it easier to reuse materials. They hope this is one effort that other competitors will copy.[2]


IBM Watson is being used inside some electronic textbooks to provide natural language, one-on-one tutoring to students on the reading material.[3]

Ashok Goel, a professor at Georgia Tech, also used Watson to create a virtual Teaching Assistant called “Jill” to reduce the workload of his teaching staff (who received over 10,000 questions per semester!). Jill was able to answer students’ questions when it had a 97% certainty of an accurate answer.[3]

Goel plans to bring Jill to the wider world of education. He expects Jill’s question-answering abilities to be an invaluable asset for massive online open courses (MOOCs) where students generally don't have the opportunity to engage with an instructor.[3]

Food and Beverage

MISO Robotics' Flippy [4]


Flippy is the world's first learning-capable robotic kitchen assistant[1]. It is installed in less than five minutes and can be integrated with point-of-sale (POS) systems. Flippy uses artificial intelligence and can learn from humans and surroundings, meaning that it can learn new skills over time. It uses cameras, sensors, and 3D and thermal scanners to see, and lasers allow staff to collaborate with it safely[2].

Moley Robotic Kitchen

Moley Robotics has created the world's first robotic kitchen. The fully-automated prototype was successfully premiered in 2015 at Hanover Messe, the international robotics show, and the consumer version is set to launch in 2018. It includes a pair of fully articulated robotic hands, oven, hob, and touchscreen unit and can be operated remotely via smartphone. The kitchen will be supported by an iTunes' style library of recipes which can be downloaded.[3]

The robotic hands are able to fully replicate the function of human hands with the same speed, sensitivity and movement. When Master Chef Tim Anderson was asked if the robot makes better lobster bisque than him, he said, "Yep. It's more reliable. It makes a good bisque every time. I'm only human—there's a lot more variation when I make it." [4]

There is also a case to be made about the educational use of a robotic kitchen wherein that you can learn how to create a complex dish by watching a pair of hands make it in front of you. By way of how the system works, you could theoretically download a Gordon Ramsay dish and have it prepared live.[5]

Moley Robotic Kitchen [6]


DoNotPay [1]


DoNotPay[1] is a chatbot that provides free legal counsel in over 1,000 legal areas[2], including parking ticket disputes, cease and desist letters, and seeking asylum. Users simply type in questions using natural language such as "I got an unfair parking ticket," and receive directions on how to proceed. DoNotPay is available in the United States, United Kingdom, and soon, Canada[3].

Unlike ROSS Intelligence, DoNotPay is unable to make its own decisions, but rather directs users through already-formulaic processes such as filing for appeals or creating disputes[4].

Criteria for Implementing RPA

1. One of the main benefits of RPA is to relieve humans of high labor intensive tasks. These tasks are one of the key issues that this technology will alleviate. [5]

2. Automating these processes with simple rules and logic will reduce the cost and reduce the time spent. There should only be a few choices for these processes and results are clear. [5]

3. Working with structured data is more efficient and is also an easy task using RPA. [5]

4. As a support tool, RPA will help staff respond to higher demand by reducing the work on the human without reducing throughput of the process. [5]

5. Similar to the previous criteria, automating simple sub-processes will help the speed up the entire process. [5]

EY RPA HR Onboarding [6]

Integration Opportunities and Evolution of RPA

Evolution of RPA

The evolution of RPA can be seen from its initial uses in simple robotics, to more advanced applications in cognitive robotics, and intelligent robotics [1]. The first applications of RPA were used in screen scraping which essentially extracts text from applications and programs in order to reformat it to be handled by other applications [1]. Screen scraping as an independent function is mostly obsolete because virtually all programs automatically perform this task. As RPA becomes more sophisticated, simple rule-based RPA applications on its own is less common because organizations are leveraging the benefits of RPA combined with newer technologies.

The next stage of RPA is seen in workflow automation tools. This application is considered the bridge between the “do” and “think” stages as workflow automation is the process of moving data, information, or tasks through automated processes that adhere to specific procedural rules [1]. The sophistication of these processes depends on the type of RPA being used, and the level of intelligence programming required to complete the task. As more capabilities are pre-programmed in that allow applications to learn, there is movement along the evolutionary graph from simple "do" tasks, to more sophisticated learning capabilities.

The final stage of RPA, and the most recent application includes RPA combined with artificial intelligence. Artificial intelligence is the capability of computer systems to perform tasks that normally require human intervention or intelligence [1]. AI is considered the “think” stage of the robotic chart and allows for more sophisticated process automation. This is becoming the most popular application as organziations are able to customize the abilities of their RPA applications, and create more effective, and efficient processes.

Milestones of RPA Evolution

Screen Scraping: The ability of these early systems to scan large sets of static information or other visual representations of data to pull key terms, integers, or other important analytics [2]. Common applications of screen scraping include human copy and pasting, text pattern matching, HTTP programming, and HTML parsing. Screen scraping is primarily used for data mining and can be seen in applications such as web indexing, web mining, online price change monitoring, and price comparison [2]. Amazon is one of the best examples to use for screenscraping as their price comparison applications are an essential aspect of their business.

Workflow Automation Tools: Workflow automation software can, for example, aid in order processing by capturing certain fields of interest, such as customer contact information, invoice total, and item ordered, translating them into your company’s database, and notifying the corresponding employee [3]. The benefits derived from workflow automation allow employees to focus on their work, rather than the underlying processes that support them. Workers are able to spend less time on repetitive tasks and more on their skills and knowledge. Critical organizational activities are able to happen seamlessly without the need to incorporate redundant processes[3]..

Artificial Intelligence: the capability of computer systems to perform tasks that normally require human intervention and intelligence. Artificial intelligence has the ability to interpret situations and their context. This allows them to form a richer understanding of situations, objects, and emotions that words relate to. These applications can already be seen through textual analysis and natural language processing that are currently in use in customer service, insurance industries, and public organizations which process large amounts of digital data, such as emails, online feedback, customer chats, and documents [4]. .

Future Integration Opportunities

There are several areas of interest concerning the future applications of RPA including integration opportunities with the internet of things, artificial intelligence, machine learning, and blockchain.

Internet of Things

The internet of things is expected to grow 31.72% each year until 2019, and expected to contribute between 10 and 15 trillion dollars to the global GDP within the next 20 years [6] . The potential for IoT with RPA includes the opportunity for improved information and analysis which results from increased number of connections between IoT devices, and business operations [6]. For example, companies could accurately track products as they physically flow through their supply chain using sensors built into physical inventory, and apply the data to greater analytics and logistics capabilities. Companies can even monitor the movements of customers as they browse physical stores using personal IoT devices. Two of the biggest challenges posed by IoT are security, and costs. Connected devices are still costly, especially for individual users and small businesses. In addition, the risk of hackers accessing private information is also a threat. Current applications of this integrative technology include Amazon’s no-checkout store, and the RFID tags used at Zara and Lululemon. The RFID tags used within Lululemon demonstrated a 98% inventory accuracy rate within the first 11 months of implementation. Lululemon employees were able to tell exactly which products were available, and their location within the store [7]. In addition, their CIO credits an increase in ecommerce sales of 8% to more effective tracking due to choosing whether certain goods should be available in-store or online based on their most profitable movement [7].

Artificial Intelligence

There are two types of intelligence that need to be differentiated: artificial intelligence and machine learning. Artificial intelligence is the broader concept of machines being able to mimic human decision making processes and carrying out tasks in ever more human ways [9]. The integration of AI algorithms into RPA offer the opportunity to improve labour efficiency and productivity accomplishing higher levels of accuracy, extend business automation to new strategic areas by automating tasks that once were reserved exclusively for humans [9]. Current applications of AI combined with RPA are customer service chat bots that combine natural language processing, with RPA processes. Many organizations have implemented these chat services as a way to increase customer service efficiency, reduce costs, and extend their service hours to 24 hours.

Machine Learning

Machine learning is the current application of AI based on the idea that we should be able to give machines access to data and let them learn for themselves [11] Machine learning is driving AI development, in that AI is becoming more intelligent and evolving from thinking within programmed parameters, to learning outside the bounds of programmed language. For example, IBM’s Watson is an intelligent machine that was pre-programmed with specific information and based on his interaction with humans and the external environment, learns to understand new information and how to more effectively answer questions [11]. By analyzing unstructured data, machine learning allows RPA to transition from doing to thinking.


Source to pay scenarios are increasingly relevant in business scenarios. By leveraging the capabilities of blockchain, RPA can be combined with the decentralized network to enable faster access to data and the creation of smart contracts [13]. For example, insurance agencies can leverage the use of sensors on cars to determine accident fault through smart contracts. A smart contract is a digitally signed computable agreement between two or more parties. When an accident occurs, claim status is then assigned to the guilty party and contracts are automatically enforced based on pre-programmed rules[13]. The role of RPA in this context is the necessity of a virtual third party (software agent) that can execute and enforce the agreement, and then use the data to facilitate faster claim resolutions. Blockchain capabilities would provide faster access to data, and allow machines with RPA capabilities to increase their efficiencies. The data transmitted through the blockchain network would allow organizations to apply RPA functions more efficiently, and utilize a more secure network.


Saba Ghodrat-Zadeh Jimmy Pham Owen Kwok Lauren Robertson


  1. 1.0 1.1 1.2 1.3 Looking Forward Looking Back, Five Key Moments in the History of RPA. (2016). UiPath. Retrieved from
  2. 2.0 2.1 Vargiu & Urru (2013). "Exploiting web scraping in a collaborative filtering- based approach to web advertising". Artificial Intelligence Research. 2 (1).
  3. 3.0 3.1 What is workflow automation (2017). Ninetex. Retrieved from:
  4. Expert System Partners Blue Prism Extend RPA Reach Cognitive Computing. (2017). BluePrism. Retrieved from:
  5. (2017). Retrieved from
  6. 6.0 6.1 RPA and the Internet of Things. (2017). UiPath. Retrieved from
  7. 7.0 7.1 Swedborg, Claire. (2016). RFID Brings Lululemon Accuracy up to 98%. RFID Journal. Retrieved from:
  8. (2017). Retrieved from
  9. 9.0 9.1 Expert System Partners Blue Prism Extend RPA Reach Cognitive Computing. (2017). BluePrism. Retrieved from:
  10. (2017). Retrieved from
  11. 11.0 11.1 Changing the World With Automation. (2017). UiPath. Retrieved from
  12. (2017). Retrieved from
  13. 13.0 13.1 TechTarget. (2017). From blockchain to RPA: A look at cutting-edge tech and the enterprise. TechTarget Network. Retrieved from:
Personal tools