Cognitive Automation and the Disruption of Business Services
By Cliff Justice, Partner, Innovation & Enterprise Solutions, KPMG
The advancing capabilities of Robotic Process Automation (RPA), Machine Learning and Cognitive Technology holds the promise of transforming Business Services and creating a new category of digital labor. But in order to avoid some of the potential pitfalls of such rapid proliferation without IT involvement, CIOs should quickly gain an understanding of what the convergence of Robotic Process Automation and Cognitive technology is, how it is used, and how best to deploy it in their enterprise.
Aligning RPA with CIO concerns
According to a 2015 Harvey Nash survey, corporate boards are asking their CIOs to focus on priorities that can increase operational efficiencies, improve business processes, cut costs, and enable business change1. Cognitive Automation offers companies numerous benefits that align with these priorities, among them, increased labor capacity and efficiency while reducing costs, the ability to refocus talent on high value business needs, and enhanced quality of service with fewer errors and standard process output.
Because of these benefits, demand for Robotic Process Automation, enhanced with Cognitive Technology is expected to increase sharply over the next three years, particularly for the functions of IT processes, sourcing and procurement, finance and accounting, and supply chain and logistics, according to a recent KPMG survey2. For their part, CIOs do recognize that significant IT transformation is on the horizon, with 9 in 10 surveyed saying they believe digital disruption will affect their organization in the next decade.3
Getting RPA off the ground
Before embarking on an enterprise-wide RPA program, executives should consider implementing a set of small pilot projects within targeted areas of the organization to acquire a better understanding of the potential opportunities RPA presents in a given environment. As these pilot programs mature and expand, CIOs can expect to better understand the materiality of benefits and potential risks within the designated scope, while at the same time gaining the experience to ready the technology and change management approach for enterprise-wide use across a wider array of business functions Many opportunities exist across transactional business activities to test and deploy RPA. Some common examples include:
• Order entry
• Accounts payable
• Approval tasks handled in Asset Management and Change Management
• Reactive and proactive Problem Management tasks
• Service Desk Support functions such as handling password resets and troubleshooting with a user
• Rules based Risk and Governance processes such as Access Management and the granting of user permissions
• Server build and pre-deployment quality assurance activities
Using a Center of Excellence approach to Create Scale and Drive Adoption
To help manage change, increase adoption,
Several of our clients have found that the successful implementation of an RPA program in one small part of the business has created a catalyst for further automation in other functions in the enterprise. In these cases, the COE has played a vital role in facilitating the expansion of automated processes, as well managing change and galvanizing support from the stakeholders.
During the initial RPA deployment, the COE is formed and focuses on gathering the knowledge and codifying the processes and mechanisms to capture and digitize process knowledge and document lessons learned. Then COE can create process libraries and toolsets that can be redeployed across other areas of the enterprise and enable subsequent programs with minimal effort.
Realizing the benefits of bots
While cognitive automation and RPA can provide a number of benefits to a business in terms of cost savings and greater efficiencies, they also present their own unique issues that CIOs must be able to address.
To be sure, the fundamental role and responsibilities of IT don’t change in an RPA environment. Development, testing, operations support, and performance tuning are all necessary components to a successful bot solution. For example, bots are considered users. For system changes, when user acceptance is identified in testing plans, testing scenarios must be devised and conducted just as they would for a human user.
Nevertheless, dealing with bots may require some adjustments. For instance, the role of QA may indeed grow. Consider this situation: A business user may request a small change to the user interface of a system to make it easier to use.
Traditionally, these changes are low risk. But where RPA is deployed, even simple user interface changes can negatively affect bot performance. Bots expect fields, text boxes, etc., on screens to be presented in the same place every time. Human users can be trained for UI changes, but bots must be reconfigured to “learn” the new screens. Additionally as bots become more efficient through machine learning, QA must ensure that the bots are learning the right resolution to issues, as those learnings have the ability to proliferate much faster.
As RPA is implemented, humans and bots will likely exist in the same workflow. But, of course, bots aren’t subject to the same time constraints as human users: They don’t need to adhere to work hours and can process work much faster. As parts of a process are sped up, and other parts remain the same, being able to predict where and to what degree bottlenecks will occur allows for better resource planning—not to mention having a tremendous positive downstream impact on the overall success and return on investments.
How can CIOs achieve success in developing and executing an RPA strategy? A key and fundamental factor is demonstrating how bots can help solve business problems and how they fit into the organization’s existing architecture/toolset.
Here are some steps CIOs should consider to ensure that the message of adding value is communicated to the business.
• Work with business partners to define use cases and proof-of-concept opportunities. Or pilot RPA in a small setting first and highlight its benefits to your business partners.
• Work with business partners to define the best scripting and support model. It comes down to a balance of “do it for them, do it with them, let them do it themselves.” Remember, as soon as a bot stops working and affects a business process, the helpdesk will hear about it.
• Have a demand-management capability before marketing RPA. Business units can get very creative with RPA and demand will pick up quickly. Have an idea of how prioritization, license allocation, hardware asset allocation, etc., will be governed.
• Form a business case for each opportunity and hold people accountable for the savings they identify related to RPA. Many of the grassroots opportunities will come from teams who have multiple people doing a task for part of their day. In order to achieve savings, work has to be re-allocated to remaining staff.