I am telling the story of how I and co-researchers built our action principles for effective automation. In this Part 2, I detail the 41 material risks we discovered—and that need to be mitigated—on the way to establishing some 30 action principles for automating effectively.
In Part 1, I outlined how our research produced some powerful action principles. But the hard truth we have chiselled out in studying emerging technologies is that no finding is for all time and for all contexts. Managerially, there is no such thing as ‘best practice’ and ‘one size fits all’. This has three consequences:
We asked a very straightforward question to make sense of the proliferation of service automation jargon: What types of tasks are the tools designed to automate or augment? Looking at the data, process, and outcomes, we conceived of service automation comprising a continuum (see Figure 1). Anchored on one side is the realm of RPA; the other, Cognitive Automation (CA). (We do not call CA technologies ‘Artificial Intelligence’ because the ‘AI’ label aggrandises what these tools, do in our opinion—I have discussed this in previous blogs.
- • Firstly, research has to be cumulative; new experiences and lessons always arise, not least because the technologies move on so fast.
- • Secondly, you have to look at all experiences—that is the full range of successes, and relative and absolute failures. Initially we looked at successes. Subsequently, as I will report below, we looked at other experiences, and from that we produced a whole book about the 41 material risks you run, and need to mitigate when automating the enterprise.
- • Thirdly, if organisations are at different stages in their capacity to manage automation, then they will need different action plans to progress.
The Research Gains Momentum
While interviews on client adoptions continued, we also collected survey data. For the past two decades, we attended the International Association of Outsourcing Professionals (IAOP)’s Outsourcing World Summit (OWS). We collect data each year in the form of a brief survey during the client-only and provider/advisor-only networking sessions. In 2015, we made service automation the topic. The surveys assessed the maturity of service automation adoption; the drivers of service automation adoption; the perceived automatability of existing business services; and the preferred sourcing option. We collected 143 completed surveys, filled out by 63 clients, 64 providers, and 16 advisors (i.e., consultants). The client respondents were senior leaders in charge of sourcing strategy, governance, procurement and provider management. Client respondents represented organisations from a variety of industries including financial services; software; technology; engineering services; manufacturing; aerospace; pharmaceuticals; life sciences; healthcare; and other industries. Provider and advisor respondents represented organisations of varying sizes and geographic locations. Our first survey found that:
- • Client organisations were increasingly expecting their outsourcing providers to help automate services. The do-it-yourself model was not their preferred route, which differed from our first four case studies.
- • Client, provider, and advisor communities believed that service automation tools could decrease costs AND improve service quality. In other words, multiple business benefits were expected, consistent with the first four cases.
We published the results in the IAOP’s publication, Pulse Magazine (Lacity et al., 2015). From the survey, the action principles table was updated (not shown in this paper) to reflect that ‘do-it-yourself’ was not the only viable sourcing model. By now, readers see the iterative process of creating action principles. As more data is collected, the list of action principles is revisited, revised and reviewed by the researchers.
Additional Field Research
During 2015 to 2017, we conducted more interviews with clients, their service automation providers and advisors, to build a total of 22 detailed client adoption journeys (see Table 1). It is useful to establish that the 22 organisations were all ‘successful’ adopters, since this relates to the relevance of the action principles we derived. We have permission to name eleven of the client organisations. We assigned pseudonyms to the other client organisations (indicated by an asterisk in Table 1). The following information can be associated with the table’s contents:
- • Eight of the client organisations are headquartered in the UK, five are based in the US, two in Germany, and one client organisation is based in each of the following countries: Australia, France, Netherlands, South Africa, Sweden, Switzerland and Russia.
- • The client organisations represent 14 industries, illustrating that service automation is not restricted to certain sectors.
- • Among our client adoption journeys, 17 adoptions were led by business operations, four were led by people from IT departments, and one was led by an innovation centre.
- • Fifteen clients adopted automation technologies that fall within the realm of RPA, three fall within the realm of CA, and two used both.
- • The 22 client organisations adopted service automation tools/platforms from Automation Anywhere (n=1); Automated Insights (n=1); Blue Prism (n=13); Celaton (n=1); Expert Systems (n=1); IBM Watson Services (n=3); IPsoft Amelia (n=1); and Redwood (n=1).
- • Five advisor organisations are represented in the study: The Everest Group, KPMG, HfS, Alsbridge (since bought by ISG), and Information Services Group (ISG).
- • In addition to these empirical methods, several providers gave product demonstrations and two authors completed an RPA foundations course to assess the claims about ease of use.
In 2017 and 2018, we also conducted multiple interviews on failures, which were largely informed by the providers and advisors. (Few clients wished to discuss failures, but some clients shared stories of first failed attempts before a successful relaunch). We asked providers and advisors to diagnose the practices which lead to failure. We documented these as the ‘risks’ shown in Table 2.
We were now not only able to assess that successful outcomes that were associated with applying action principles, we could now show that failures often were associated with not applying action principles.
After the first survey at the IAOP OWS in 2015, we collected surveys at the OWS in 2016, 2017, and 2018. The surveys are not very rich, but they do indicate trends, at least for the IAOP community. The surveys showed increasing rates of RPA and CA adoption in client and provider firms, increased evidence for multiple business benefits expected and delivered, and varied approaches to service automation sourcing (do-it-yourself, rely on the provider, engage an advisor, buy as a service). (Lacity et al. 2015, 2016, 2017a, 2018b).
In 2018, we decided to assess the action principles with a more thorough survey of client adopters. Partnering with Dr John Hindle of Knowledge Capital Partners (KCP) and Dr Shaji Khan from the University of Missouri, we developed a survey with 48 detailed questions based on our action principles. We received 764 responses, but most clients were not far enough in their adoption journeys to report outcomes and dropped out of the survey midway. We did get 112 completed surveys that reported on 238 RPA deployments. Once again, clients reported multiple business outcomes, including positive returns on investment, increased business agility and compliance, and enhanced customer experiences. Interestingly, this survey contradicted the IAOP survey in one regard: Relying on a traditional BPO provider to lead the service automation was ranked the lowest in terms of business value delivered in the KCP survey; it was rated the among the best sourcing approaches in the IAOP surveys. We divided the KCP results over five reports because we discovered so many complex relationships between action principles and outcomes (Hindle et al. 2018abcde).
In summary, by 2018, we had gathered rich data from the 22 detailed successful client adoption journeys, multiple interviews into dozens of failures, and five surveys. From the information we were able to formulate the RPA Risks and Risk Mitigation Framework shown in Table 3.
We had produced a highly impactful set of findings, and then intelligent automation began to happen, and the technology continued to progress.
In Part 3—in next month’s newsletter—we report on researching intelligent automation cases, and how we developed further action principles for the emerging automation realities.
In case you want to follow up in detail on any part of our automation studies, I list below the relevant publications that cover the period 2015–2021:
Lacity, M. (2017). Reimagining Professional Services with Cognitive Technologies at KPMG. UMSL Working Paper, http://www.umsl.edu/~lacitym/LacityKMPGFinal2017.pdf
Lacity, M., Babin, R., & Willcocks, L. (2017). Research Center: Service Automation Trends Survey. Pulse Magazine, Issue 28, 40-44.
Lacity, M. & Willcocks, L. (2015). What Knowledge Workers Stand to Gain from Automation. Harvard Business Review Online, https://hbr.org/2015/06/what-knowledge-workers-stand-to-gain-from-automation
Lacity, M. & Willcocks, L. (2016a). Speed of Automation Adoption Faster for Providers than Customers. Pulse Magazine, May/June, 10-17.
Lacity, M., & Willcocks, L. (2016b). A New Approach to Automating Services. Sloan Management Review, 57(1), 41-49.
Lacity, M., & Willcocks, L. (2016c). Robotic Process Automation at Telefónica O2. MIS Quarterly Executive, 15(1), 21-35.
Lacity, M. & Willcocks, L. (2017). Robotic Process Automation and Risk Mitigation: The Definitive Guide, Stratford-Upon-Avon, SB Publishing.
Lacity, M. & Willcocks, L. (2018a). Robotic Process and Cognitive Automation: The Next Phase, Stratford-Upon-Avon, SB Publishing.
Lacity, M. & Willcocks, L. (2018b). OWS18 Findings Revealed on Client Service Automation Deployments. Pulse Magazine, https://iaoppulse.net/2018/06/research-corner-ows18-findings-revealed-on-client-service-automation-deployments-what-do-they-mean-for-your-job-and-organization/
Lacity, M., Willcocks, L. & Craig, A. (2017). “Service Automation: Cognitive Virtual Agents at SEB Bank,” The LSE Outsourcing Unit Working Research Paper Series.
Lacity, M. Willcocks, L. & Gozman, D. (2021). Influencing information systems practice: The action principles approach applied to robotic process and cognitive automation. Journal Of Information Technology, 36, 3, 216-240.
Lacity, M., Willcocks, L., & Gunkel, G (2017). Smart Sourcing: Cognitive Automation at Zurich Insurance. Intelligent Sourcing Magazine, Issue 4, 34-42.
Lacity, M., Willcocks, L., & Yan, A. (2015), Are the robots really coming? Service Automation Survey Findings. Pulse Magazine, Issue 17, pp. 14-21.
Lee, A. (1999). Rigor and Relevance in MIS Research: Beyond the approach of positivism alone, MIS Quarterly 23(1): 29–33.
Willcocks, L. & Lacity, M. (2016). Service Automation: Robots and the Future of Work. Stratford-Upon-Avon, SB Publishing.
Willcocks, L., Hindle, J. & Lacity, M. (2019) Becoming Strategic With Robotic Process Automation. SB Publishing: Stratford.