By Jean-Jacques Bernard and Hans Castel, Oracle Insight
For organizations with shared service centers (SSCs), decreasing service costs and response times are among the highest priorities. In a 2017 Deloitte report, 73% of respondents reported shared services productivity increases of 5% or more year over year. However, due to the law of diminishing marginal productivity, it will become increasingly difficult for SSCs to continue achieving that goal without a fundamental shift in how they operate.
That transformation can be brought about with robotic process automation (RPA) and artificial intelligence (AI). These emerging technologies can help increase productivity levels without compromising quality in SSC operations. RPA is the automation of routine tasks in office environments. It focuses on automating tasks such as file handling, data entry, and other necessary but no-added-value activities to improve overall productivity. AI leverages analytical science and advanced machine learning algorithms to analyze internal and external information in order to draw insights. It has the potential to improve productivity and ease decision-making.
While business adoption of AI is still in early stages, with only one in five organizations having incorporated AI in some process according to an MIT Sloan Management Review report, the adoption of RPA to replace more-simple routine tasks is increasing fast, with a global annual market growth of circa 64%.
Organizations need to prepare their workforce for the massive changes that AI will bring.
Both technologies involve potential pitfalls for organizations whose leaders rush to adopt them without due preparation and assessment of the human element. The pitfalls include unclear responsibilities, lack of focus, and bad data or lack of data. Leaders should create a strategic roadmap and take a step-wise approach in introducing these new capabilities.
The following strategies can help SSCs successfully adopt RPA and AI:
1. Start small. Deploying RPA is a logical first step in optimizing shared services, because it has a quick return on investment. We expect that companies will integrate RPA capabilities into IT platforms or leverage them as a service via cloud, especially in enterprise resource planning (ERP)–related functional areas. Start with RPA to automate simple processes such as file merges and file uploads. Increase complexity as the organization matures its capability to implement RPA. An example of process improvement is reported by AFP, where a robotic cash pooling program replaced the six-to-eight hours of manual effort required to reconcile the cash pool balances with an automated process that took just five minutes. At the same, the process ran more frequently, improving the quality of the outcome. Given that SSCs can implement RPA in a relatively short amount of time, they can potentially achieve a return on investment within just months.
2. Aim big. For more-complex processes, explore AI. AI can help executives derive meaningful insights from internal and external data and provide a course of action. Use cases that are emerging include automated expense reporting, where companies can achieve close to 100% of compliance checking and fraud detection. To determine where AI should be tested, companies need to focus on areas in which cognitive tasks are difficult to automate. Listing all processes that demand analysis is a way to identify AI candidates. The only constraint is to ensure data is available to be able to train the AI algorithms.
3. Ensure your organization is ready for cross-functional collaboration. Both RPA and AI will challenge organizational boundaries, so cross-disciplinary teams will be needed to implement the technologies. These teams need to be empowered to make decisions on their own and act on them, so they will need to have high-level sponsors in the organization that enable them to work without constraints.
4. Take a hard look at your IT capabilities. Not all IT systems will be able to support RPA and AI. Ease of integration, strong data governance, and data access are critical for success. Without them, RPA and AI initiatives won’t be able to deliver the expected benefits. In addition, organizations will need to implement some of the capabilities into core management systems, such as ERP or human capital management systems, to enable quick implementation in SSCs. (One example is using RPA to detect fraud in expense reports, which are in the core ERP system.) In this regard, software as a service (SaaS) is the way to go for implementing AI on standardized data, given its ability to quickly deploy new technologies.
5. Assess the impact on people and plan training early. Organizations need to prepare their workforce for the massive changes that AI will bring. AI will either further automate complex tasks or provide probable outcomes to facilitate decision-making. With all simple tasks automated, the remaining nonautomated tasks and decision-making will focus on the most difficult processes, which will require stepping up the capabilities of the workforce that is working with AI. Beyond training, organizations will need their subject matter experts to thoroughly explore the output possibilities of the algorithm they deploy to ensure proper use of the results.
SSCs are already seeing significant efficiency benefits as they deploy RPA and AI. Case studies suggest that in some cases, 80% of human-involved automated tasks in SSCs can be replaced. The released workers can be redeployed to higher value-add activities. This helps SSCs drive strategies to expand their delivery model and take on new tasks. At the same time, RPA and AI will drive process improvement. With these new capabilities, companies can enlarge the scope and reach of an SSC and achieve additional value beyond cost optimization.
Photography by Julian Santacruz