Laurent - stock.adobe.com
CGS is making robotic process automation the key vehicle for digital transformation, tapping UiPath RPA software and building a center of excellence to support companywide adoption.
The New York-based company provides business process outsourcing (BPO) services, a retail industry software suite and IT outsourcing offerings, such as managed IT services. CGS employs more than 7,500 people, providing services to customers in 45 countries.
John Samuel, executive vice president at CGS, is the company's digital transformation advocate. "When we started our journey, we looked at technologies and process automation areas that would be the most impactful to the organization, and one of the first areas … was RPA," he said.
CGS launched the RPA rollout in its BPO division, the company's largest business segment and employer of thousands of call center agents. The company's other divisions and geographic operations followed. Companywide RPA deployment began in the first quarter of 2019.
Samuel said CGS tapped UiPath as its RPA software provider after researching vendors in the RPA market. He said many considerations informed the selection, but UiPath's aggressive technology roadmap was the deciding factor.
"The primary [factor] is the pace at which they are innovating the technology and extending it outside of the traditional robotic process automation framework," he said. Specifically, Samuel cited UiPath's push to move the platform from RPA to intelligent process automation, or IPA. IPA aims to extend RPA with cognitive technologies.
Samuel said UiPath has built cognitive services into its RPA platform, and he also noted the ecosystem surrounding the platform, which also contributes to the IPA direction. He said CGS formed a global partnership with UiPath earlier this year.
Departing from traditional RPA
The CGS deployment differs from traditional RPA rollouts in a couple of ways. Some organizations have counted the number of software bots deployed to chart progress. But CGS focuses on the number of processes executed as a more meaningful measure of RPA adoption.
Samuel said CGS uses a mix of attended and unattended software robots. The unattended robots function in a concurrent execution model, which makes it hard to come up with an exact bot count, he noted. A small number of bots licensed to run concurrently could execute numerous processes, making the actual number of bots in production an ineffective indicator of scale.
"I believe this is going to be a trend … counting the processes executed rather than the number of bots," Samuel said.
John SamuelExecutive vice president, CGS
In another departure, CGS looks beyond traditional RPA use cases, which emphasize high-volume, highly repetitive processes. Samuel said, when he is stocking the company's bot project pipeline, he looks for not only RPA's low-hanging fruit, but opportunities for addressing quality issues.
For example, a process spanning several systems might involve multiple data entry tasks, which could introduce human error. Quality issues in such a setting could expose an organization to reputational risk, Samuel said.
"Even if the transaction volume is not there … there may be areas with potential quality issues," he said. "Those are great targets for us."
CGS also looks for use cases where software robots can support compliance activities or provide greater process consistency. As for the latter, several employees may be assigned to execute a complex process and may do so in varying ways. A software bot, in contrast, will execute a process consistently every time, Samuel said.
The company also looks to employ robots where speed is a consideration. Cost efficiency, meanwhile, is not a primary RPA driver but rather a byproduct of other RPA benefits, Samuel noted.
"If we can find a use case where we can check the boxes for every one of our criteria, that goes to the top of the list immediately," Samuel said.
Ideas to consider when adopting RPA
Samuel mentioned a few elements of his company's RPA pursuit.
Educating an organization on what RPA is and what it does is critically important.
"There is an assumption that everyone should know some of the buzzwords, and our industry adopts certain terms, but they are not necessarily clear to the business community," he said. "Education is really key to adopting these types of technologies."
Agile and MVP
Samuels said Agile methods and the concept of minimum viable product are part of the company's RPA adoption. "This doesn't have to be a project where it needs to take a long time," he said of RPA. "These are very quick-win types of projects if they are done right. The key is to do pilots and refine them and embrace a continuous improvement process."
CGS looks to deconstruct a process into smaller units. "When we evaluate a process, instead of taking and building a massive process into an RPA, we break it down into subprocesses," Samuel said. The smaller processes are easier to execute and troubleshoot, he noted.
A smaller, discrete process may be applicable to a range of RPA projects across the organization -- a module that logs in to a Salesforce platform, for instance. Such processes may be housed in a repository for reuse, Samuel noted. UiPath RPA software provides libraries for this purpose. CGS can also take advantage of prebuilt processes available in UiPath Go, the RPA vendor's online marketplace.
Sizing the environment
RPA software vendors often provide calculators to determine how many unattended and attended bots will be needed and estimate the duration of work performed and the quantity of execution events. Organizations equipped with this data can create the appropriate number of bots and monitor contention for execution. RPA adopters can also assess the nature of tasks to determine the best RPA licensing approach. For example, a lower priority task that is not time sensitive can wait until a bot is free to execute the task, Samuel noted. That type of task would be a candidate for concurrent licensing. Conversely, a task that must be executed and is time sensitive may not be a great candidate for concurrent licensing, he added.
At CGS, the company's RPA philosophy revolves around decentralized teams that adhere to centralized rules. The local teams brainstorm ideas for RPA projects and power users within those teams to build the bots. The teams bubble up organically within the company's divisions and geographic areas.
"Instead of hiring an army, find the people who are going to be passionate across the organization, and build a cross-functional team," Samuel advised.
CGS has put a corporate RPA governance program in place, creating a centralized center of excellence to provide guidance on how to adopt and deploy RPA.
"If you don't create some structures around this, you are going to have standards, processes and methods being developed across the organization globally that are going to deviate," Samuel said. "You are going to have to deal with cleanup and strategic alignment down the road."
Free-form RPA initiatives risk the cost of duplicated effort. And Samuel pointed out that a local group may forget to think about the ongoing maintenance and management of bots.
"It makes sense to have a structure that mandates certain operational requirements," Samuel said.
Power users, for example, should follow standard practices to avoid pitfalls such as embedding credentials into the automated scripts they are building. The bot builders should also make sure their software goes through a proper testing process, he added.
The centralized standards and processes are designed to make bots safe, effective and maintainable. But local teams still have the flexibility to depart from the guidelines where it makes sense, Samuel noted.
"Local teams are encouraged to do development and assist in the process but follow corporate guidelines and best practices," he said.
In addition, the company's centralized resources "facilitate dialogue between local teams and foster an environment of sharing.," according to Samuel.
From RPA to IPA
CGS aims to socialize the capabilities of RPA to the organization and "getting that to be the prominent focus on the digital transformation journey," Samuel noted.
Beyond that focus, the company's RPA advocates will look for processes that have not been great candidates for RPA in the past because of their complexity. He said the task is to find ways to incorporate computer vision and cognitive services, such as natural language processing, machine learning and other AI-centric services, into workflows so decisions can be made without waiting for human mediation.
CGS plans to push the RPA boundaries with R&D projects where the company can experiment on processes that are more complex than its current crop of software bot use cases, according to Samuel.
"We are moving to intelligent process automation," he said.