JIS Workshop on Robotic Process Automation (RPA) Research: Views from RPA Industry Leaders and AIS Researchers

Workshop Description and Objectives

Panelists

References

Workshop Recording

Workshop Q & A


Workshop Description and Objectives:

The aim of this workshop is to bring together Accounting Information Systems (AIS) researchers, representatives from leading RPA product and service providers, RPA users and practitioners working with or considering working with RPA, to have a conversation on RPA related challenges and research opportunities.

We hope through the interaction to achieve the following objectives: First, provide an opportunity for RPA software, implementation firms and users to articulate some of the issues/challenges they face. Second, provide an opportunity to researchers interested in these topics/ideas to start exploring the possibility of collaboration with practitioners on addressing these research questions. Third, convert these research topics/ideas into a call for papers for a special issue of JIS on the topic of RPA in order to expand our understanding of this growing area of activity.

Background:

The IEEE Corporate Advisory Group (IEEE 2017) describes robotic process automation (RPA) as an instance of preconfigured software that uses a combination of business rules and predefined “activity choreography” to complete a series of tasks on one or more software systems. In simple terms this means that while RPA is like recorded macros in Excel, it can be configured to perform tasks based on different systems. For example, it can read emails and trigger responses, open PDFs and use OCR to read contracts, login into ERP systems to read/write data, leverage human-supervised machine learning (ML) to identify non-explicit rules, or enable customers to interact directly with systems via chatbots (Moffitt et al. 2018; Plattfaut 2019; Miers et al. 2019).

State of Adoption:

Approaching the RPA from both the supply side (firms involved in the delivery of RPA solutions) and demand side (firms investing in RPA initiatives) there is ample evidence that the market is vibrant. According to Gartner’s 2019 report, RPA is the fastest growing software segment with an annual growth of 63% (Miers et al. 2019). A 2019 survey (Protiviti 2019) has shown that relatively large companies (i.e., companies with sales of a billion dollars or more) have been investing between $10-$20 million per year in RPA projects and this is likely to continue in the near future. Respondents in the Protiviti (2019) survey reported that this spending is motivated by expected benefits in increased productivity, better product quality, and strong competitive market position. However, the success of these investments is hindered by inability to prioritize potential projects and pursue the best applications.

Challenges:

A Gartner report (Miers et al. 2019) focuses on several misconceptions and challenges associated with RPA; for example: 1) RPA adoption is not likely to lead to labor cost reduction, because the firms tend to refocus employees towards more value adding activities. Cooper et al. (2019) corroborate this based on feedback from accounting firms. 2) RPA is not suitable for automating lengthy processes. It is more suitable for processes that last just a couple of seconds. For lengthy processes, intelligent business process management systems (iBPMS) are more appropriate. 3) RPA is just one tool, rather than the only tool, in the automation toolbox. RPA should be seen as a complement of other tools as iBPMS and process mining. 4) Use of RPA has been described as “long-term technical debt” as it binds the organization to legacy systems.  5)  While some vendors claim that that their software contains machine learning (ML) algorithms, there is relatively little room for leveraging ML in the core of RPA.

Research Opportunities:

The state of diffusion of an emerging technology determines “what types of research questions we can address, what research data is available and what methodologies are available to study a technology and its uses” (O’Leary 2008, 241). The 2019 Gartner Hype Cycle does not include RPA in its list of technologies tracked by Gartner.[1] However, in a local version of the Hype Cycle (Dubai, AEE), RPA software is shown in the trough of disillusionment.[2] O’Leary argues that during this stage, information about the technology is likely to be limited. Therefore, research would mostly focus on implementation issues and limitations of the technology.

Research at the intersection of RPA and accounting has been relatively active. For example, Moffit et al., (2018) have examined the use of RPA in audit and suggested several research topics, including challenges of RPA implementation and how to address them. The study by Kokina and Blanchette (2019) leverage feedback from RPA adopters to analyze implementation issues and performance outcomes. Organizations benefit from automating rules-based, well-structured processes with digital inputs. The primary benefits include cost savings, improved process documentation, lower error rates, more accurate measurement of process performance, and better report quality. Two recent studies (Cooper et al. 2019; Cooper et al. 2020) have examined RPA implementations among the Big 4 accounting firms. While the former looked at the RPA leaders within the accounting firms to gain insights into how RPA software is currently being used, the latter examined how both firm leaders and lower-level employees perceive and respond to RPA. While both groups agree on the positive effects of RPA on employee performance and career prospects, only firm leaders believe RPA will improve work satisfaction.

Format:

  • The workshop will be held on Friday, September 25 from 1-4 pm.
  • It will be offered through the University of Waterloo over Webex.
  • Advance registration is required. Registration will open soon. Until the registration page is open please send us an email indicating your interest so that we have it in our records.
  • The registration fee is $50 but a 50% discount is available for AAA AIS Section members, AAA SET Section members, ISACA Toronto Chapter members, IIA Canada/Toronto Chapter members and CPA Canada members and a 100% discount is available for students.
  • This workshop qualifies for 3.5 hrs of CPE credit.

Structure-Timeline of the Workshop: 

  • Welcome – Efrim Boritz on behalf of UWCISA and JIS
  • Introduction by Theo Stratopoulos (10 minutes)
  • Presentations by industry speakers on products, services and issues/challenges including Q&A. (90 minutes)
  • Break (10 minutes)
  • Presentations by academics who have done RPA research including Q&A. (60 minutes)
  • Closing panel discussion and Q&A (10 minutes)

Back to top


Industry Panel:

Academic Panel:

Back to Top


References

Cooper, L. A., D. K. Holderness, T. L. Sorensen, and D. A. Wood. 2019. Robotic Process Automation in Public Accounting. Accounting Horizons 33 (4): 15–35.

Cooper, L., K. Holderness, T. Sorensen, and D. A. Wood. 2020. Perceptions of Robotic Process Automation in Public Accounting. SSRN Scholarly Paper. Rochester, NY: Social Science Research Network.

IEEE. 2017. IEEE Guide for Terms and Concepts in Intelligent Process Automation. Corporate Advisory Group. IEEE.

Kokina, J., and S. Blanchette. 2019. Early evidence of digital labor in accounting: Innovation with Robotic Process Automation. International Journal of Accounting Information Systems 35: 100431.

Miers, D., M. Kerremans, S. Ray, and C. Tornbohm. 2019. Magic Quadrant for Robotic Process Automation Software. Gartner.

Moffitt, K. C., A. M. Rozario, and M. A. Vasarhelyi. 2018. Editorial Robotic Process Automation for Auditing. Journal of Emerging Technologies in Accounting 15 (1): 1–10.

O’Leary, D. E. 2008. Gartner’s hype cycle and information system research issues. International Journal of Accounting Information Systems 9 (4): 240–252.

Plattfaut, R. 2019. Robotic Process Automation – Process Optimization on Steroids? In ICIS 2019 Proceedings. Munich.

Protiviti. 2019. Taking RPA to the Next Level. Survey. 2019 Robotic Process Automation Survey. Menlo Park, CA: Protiviti Inc.


[1] https://www.gartner.com/smarterwithgartner/5-trends-appear-on-the-gartner-hype-cycle-for-emerging-technologies-2019/

Back to Top


Workshop Recording

The recording of this workshop held September 25, 2020 is only available to registrants and requires a password:

https://uwaterloo.webex.com/uwaterloo/lsr.php?RCID=80b977d198cb4897a2bd1f72648a2ad6

Back to Top


Workshop Q & A

Session number:  1728518820

Date:  September 25, 2020

Starting time:  12:59 PM

________________________________________________________________

‑­Myles Stern­ (­m.stern@wayne.edu­)­ - 1:18 PM­

Q: ­The definition of RPA seems overly broad to me. Do you agree?­

               ‑­Ralf Plattfaut­­­­ - 1:23 PM­

               A: ­To identify the right definition of RPA is an interesting task. I agree that the IEEE definition is really broad.­‑

               ‑­Abigail Zhang­­­­ - 1:23 PM­

               A: ­I agree. I prefer referring to RPA as the Realm of RPA. RPA essentially acts an overlay ­‑________________________________________________________________

September 25, 2020             1:25 PM               from Efrim Boritz to David Wood:          Attendees are wondering whether you are willing to share your RPA illustration for teaching purposes.

September 25, 2020             1:26 PM               from David Wood:       Yes. It is also available on the EY ARC.

________________________________________________________________

‑­Jodie lobana­ (­lobanaj@mcmaster.ca­)­ - 1:50 PM­

Q: ­Jacob, what RPA software have you used? Also, how long did it take for you to learn the software?­

               ‑­Jacob Summers­­­­ - 1:51 PM­

A: ­I know how to use UiPath and Automation Anywhere. I would say I felt comfortable automating easy processes after 30 hours (UiPath has a great level 1 learning course). But I learned a lot more from actual bot building and learning from other really good developers -- this knowledge took a good amount of time in order make a given bot run more reliably and effectively.

               ‑­Abigail Zhang­­­­ - 2:06 PM­

               A: ­Depending on the complexity of business process you want to automate. But for easy ones, a 9-hour session is enough to go­‑

________________________________________________________________

‑­Jing Lu­ (­jlu11@uoguelph.ca­)­ - 2:03 PM­

Q: ­What skills do you need to build bots? Programing skills? Business knowledge? Or use the bots software such as automation anywhere? ­

               ‑­Abigail Zhang­­­­ - 2:05 PM­

               A: ­Based on my experience of self-learning and teaching RPA, learners can start with no programming knowledge. But learners will learns skills required gradually: basic programming and business process understanding­‑

               ‑­Abigail Zhang­­­­ - 2:07 PM­

               A: ­Both Automation Anywhere and UiPath provide excellent tutorials and they offer certificates once you go through their course modules and pass the test­‑

________________________________________________________________

‑­Jing Lu­ (­jlu11@uoguelph.ca­)­ - 2:07 PM­

Q: ­Thanks Abigail. Just wonder, which programming language do you use? Python? C?­

               ‑­Abigail Zhang­­­­ - 2:09 PM­

               A: ­RPA has been made almost code-fee. In other words, you can drag and drop activities to configure automation. No specific programming language is required. What's important is the programming logic.­‑

               ‑­Abigail Zhang­­­­ - 2:09 PM­

               A: ­For some complicated automation, you can embed Python or VBA in RPA bots­‑

               ‑­David Wood­­­­ - 2:11 PM­

               A: ­Jing, each software is built on a different programming platform. So UIPath is built on .net and C#. These softwares are designed to be low-code, no-code. Understanding basics of variables, if/then logic, looping, helps a lot, but is not required.­‑

               ‑­Abigail Zhang­­­­ - 2:13 PM­

               A: ­https://drive.google.com/file/d/1qhlwRh36xmcs9m2H_NcDpA-4rgLQ0eUH/view?usp=sharing Hope this video can illustrate how to "configure " rpa­‑

               ‑­Ralf Plattfaut­­­­ - 2:15 PM­

               A: ­But RPA is essentially low- to no-code­‑

________________________________________________________________

‑­Jing Lu­ (­jlu11@uoguelph.ca­)­ - 2:16 PM­

Q: ­Thanks Abigail and David. It sounds like a very promising technology. ­

               ‑­Abigail Zhang­­­­ - 2:26 PM­

               A: ­RPA is suitable for "swivel chair" tasks where you move data from one application to another. For example, if you need to get something from ERP and then process that data and put it in another application like email. ­‑

               ‑­Abigail Zhang­­­­ - 2:26 PM­

               A: ­Another situation where RPA is suitable is that ERP or other enterprise systems don't have APIs­‑

________________________________________________________________

‑­Myles Stern­ (­m.stern@wayne.edu­)­ - 2:23 PM­

Q: ­There are many different approaches to automation. For instance, most ERP systems have workflow capabilities where you can trigger an external process. How do you determine if RPA is a good fit in a specific situation. ­

               ‑­Abigail Zhang­­­­ - 2:27 PM­

               A: ­ RPA is suitable for "swivel chair" tasks where you move data from one application to another. For example, if you need to get something from ERP and then process that data and put it in another application like email. ­‑

               ‑­Abigail Zhang­­­­ - 2:27 PM­

               A: ­ Another situation where RPA is suitable is that your ERP or other enterprise systems don't have APIs­‑

               ‑­Abigail Zhang­­­­ - 2:29 PM­

               A: ­That's why it's better to think of RPA as an overlay­‑

________________________________________________________________

‑­Akhilesh Chandra­ (­a20cn16@gmail.com­)­ - 2:34 PM­

Q: ­is there a use-case/evidence/motivation for RPA in governments where politics can run a mock with other ramifications­

               ‑­Abigail Zhang­­­­ - 2:34 PM­

               A: ­https://www.uipath.com/blog/us-government-continues-paving-way-for-rpa­‑

________________________________________________________________

‑­Robert Parker­ (­Robertgparker@shaw.ca­)­ - 2:32 PM­

Q: ­In selecting business process that would benefit from RPA how  is privacy considered  ­

               ‑­Abigail Zhang­­­­ - 2:39 PM­

               A: ­It's a great question on the data privacy. Currently the software vendors have some solution on protecting the data privacy. However, I don't know how effective they are.  ­‑

________________________________________________________________

‑­Gregory Kogan­ (­grkogan@gmail.com­)­ - 2:38 PM­

Q: ­From a SOX compliance perspective it seems like RPA doesn't pose a lot of new risks, but what do you see from other intelligent automation techniques like the self-service automation work flow tools?­

               ‑­Abigail Zhang­­­­ - 2:42 PM­

               A: ­Great question. RPA follows deterministic rules and is non-invasive. However AI technologies may impose more issues. So a robot’s governance model should be in place and audit of these tools will be required­‑

________________________________________________________________

‑­Robert Parker­ (­Robertgparker@shaw.ca­)­ - 2:34 PM­

Q: ­In designing RPA how is privacy or other legal requirements considered?­

               ‑­Ralf Plattfaut­­­­ - 2:42 PM­

               A: ­As the bots often run internally using the same access rights as human users do, data privacy is typically not that much of a concern. Other legal requirements depend of course on your legal frame. ­‑

               ‑­Ralf Plattfaut­­­­ - 2:43 PM­

               A: ­In regulated industries in Germany, e.g., you need to announce RPA usage to your regulator. But usually this is not that much of a problem.­‑

________________________________________________________________

‑­Jelil Akindele­ (­b.akindele@hotmail.com­)­ - 2:46 PM­

Q: ­What are the potential risks that might deter organizations from deploying RPA?­

               ‑­Ralf Plattfaut­­­­ - 2:47 PM­

               A: ­1) Bots break if core systems change (unless you build them really well)­‑

               ‑­Ralf Plattfaut­­­­ - 2:48 PM­

               A: ­2) you add architectural complexity - just a new bot to your it architecture­‑

               ‑­Ralf Plattfaut­­­­ - 2:48 PM­

               A: ­3) you delay core system change - essentially you keep the lights on on your legacy systems­‑

               ‑­Abigail Zhang­­­­ - 2:49 PM­

               A: ­Based on my experience, some of the concerns are the mulfunction of the bot due to process changes and software updates; errors in the bot; data security (I heard one case on the bot wrongly handled the data)­‑

________________________________________________________________

‑­Akhilesh Chandra­ (­a20cn16@gmail.com­)­ - 2:49 PM­

Q: ­how does RPA affect control frameworks, if at all..­

‑­Akhilesh Chandra­ (­a20cn16@gmail.com­)­ - 2:49 PM­

Q: ­so, can BOTS be tried as humans in case of crime in the future­

               ‑­Ralf Plattfaut­­­­ - 2:50 PM­

               A: ­I guess this does not hold for bots yet, as they follow algorithms and rules - but maybe to cognitive automation and ai­‑

________________________________________________________________

‑­Jelil Akindele­ (­b.akindele@hotmail.com­)­ - 2:49 PM­

Q: ­What are the potential risks that can deter organizations from deploying RPA?­

________________________________________________________________

‑­Adam Presslee­ (­capressl@uwaterloo.ca­)­ - 1:33 PM­

Q: ­Can you discuss the role of explainability, and how you go about explaining what (and how) the RPA is doing? What are some barriers to explainability?­

               ‑­Abigail Zhang­­­­ - 3:21 PM­

               A: ­Compared to AI, RPA is considered as explainable because it follows pre-defined rules and you can see a flowchart of how the bot works. What generates concern is how to explain the functioning of AI technologies­‑

________________________________________________________________

‑­Rajendra Srivastava­ (­rsrivastava@ku.edu­)­ - 3:14 PM­

Q: ­RPA for financial audits and internal control audits, I do not see any discussion on evidential structure and various assertions. The audit evidence forms a network of variables that needs to be evaluated and combined to form an overall judgment . Imp issue­

________________________________________________________________

‑­Jonathan Ye­ (­jonathan.ye@uwaterloo.ca­)­ - 3:50 PM­

Q: ­Do you think if technology readiness affects the RPA adoption­

________________________________________________________________

‑­Deniz Appelbaum­ (­appelbaumd@montclair.edu­)­ - 3:51 PM­

Q: ­Hi David, Please share your horror stories about publishing RPA! ­

Back to Top

________________________________________________________________