Digital Twins: A New Dimension to How BPOs Can Examine Their Identity


SOURCE: NEARSHOREAMERICAS.COM
MAR 29, 2022

In the minds of most, the concept of digital twins conjures up visions of heavy industry; mile-long manufacturing plants full of large one-armed robots piecing together chassis, or a gas refinery and its bewildering labyrinth of pipes and tubes.

But digital twins aren’t only for use for large-scale physical structures. More and more frequently, digital twins are being used to create a visual representation of the processes of an organization, like a BPO. Though not so much, it seems, in the Nearshore.

“The term digital twin is much more accepted in Europe than it is in the US,” John Vaughn, Chief Operating Officer at BusinessOptix, a company that provides cloud-based digital twin capabilities, told Nearshore Americas. “And we find that operational people have a tendency to stay away as it’s considered a manufacturing or engineering term,” he added.

But with increased digitalization of the BPO landscape, including the evolution of chatbots, and the potential of tech solutions like artificial intelligence and machine learning within an organization, it might be time for the BPO companies to embrace their digital replicas.

Building a Digital Twin

So what exactly is a digital twin? IBM offers this definition: “A digital twin is a virtual representation of an object or system that spans its lifecycle, is updated from real-time data, and uses simulation, machine learning and reasoning to help decision-making.”

Effectively, a digital twin is an exact virtual model of an asset or process, and was pioneered by NASA in the early 2000s. While a digital counterpart to a factory line is straightforward to understand, mapping in 3D a process that involves the variance of human operators simply seems far more complex. After all, engineers already know how the assembly line of an auto factory should function.

John Vaughn, COO at BusinessOptix

“Digital twins are easily understood in manufacturing and product development. But when you move to an organization, it isn’t markedly different,” Vaughn explained. “What does change, however, are the variables. There won’t be much of a variance in performance between a machine that has been 9 months on the job and one that was installed yesterday. But there’s a huge difference in the ability of a BPO agent with 9 months experience against a new recruit.”

For this reason, the life of the BPO digital twin begins by collecting data. Mounds of metadata are scraped from inside of the organization, from surveys in which human workers provide step-by-step explanations of their undertakings for each task, whether making a sales cold call, receiving a customer complaint in social media, or providing expert technical support. This acquired data is tallied alongside the organization’s own structural data, including the role that any process automation solutions the company uses has. All of these thousands of pathways that every customer interaction could move along, and the communication stages and nodes through which they pass, are mapped within the digital twin’s visual ‘self’.

“The first stage of digital twin modeling we call descriptive, and that’s when the processes are mapped – what we think they look like. That’s a good start and it gets everyone together,” Vaughn explained. “Then we move from that process diagram to a 3D process model that doesn’t only include what’s happening horizontally but also along each step of the way. This includes metadata like arrival time, wait time, work time, and maps all the individuals involved in every interaction.”

“Why do wait times spike on certain days of the month, or certain times of the day? Why does a process only happen on the third Thursday of every month? Why does my office in Chicago do this when the Johannesburg office does that? Is this the same process or an acceptable regional variation?” — John Vaughn

“This stage is about creating that understanding of your resources, whether they be human or digital workers, and how they come together,” Vaughn added. “By doing this, anomalies are viewed more easily. Why do wait times spike on certain days of the month, or certain times of the day? Why does a process only happen on the third Thursday of every month? Why does my office in Chicago do this when the Johannesburg office does that? Is this the same process or an acceptable regional variation? These areas can then be investigated,” he added.

The Pareto Principle, also known as the 80/20 rule, is often used by companies looking for smoother business processes. The rule states that ‘in any given situation, 80% of the outcome is produced by 20% of the input.’ The Six Sigma and Lean techniques often employ this rule to improve business processes. And while it certainly has its use, it isn’t exacting enough to deliver the incremental improvements that deliver success, Vaughn suggested.

“A company’s risk is in the tail of that. It’s the exceptions happening that they really need to know about,” he said.

“Part of the usefulness of a digital twin lies in its capacity to be modeled on reality but developed in isolation from the real world until it approaches the best possible performance in its current and anticipated circumstances.” — Campgemini

While processes like Six Sigma look at the core systems of a business, they don’t consider all of the workflow that is happening outside of core systems, like informal emails going back and between employees, the use of a basic Excel spreadsheet or a forgotten Access database. Through process mining, a digital twin can demonstrate all of these anomalies too.

Once the informative model has been created, the digital twin can then become predictive. This includes running scenarios that provide insight into how small chances along the lifecycle of an incoming call can change wait times, CSAT, or even employee engagement scores.

As Capgemini explains, “Part of the usefulness of a digital twin lies in its capacity to be modeled on reality but developed in isolation from the real world until it approaches the best possible performance in its current and anticipated circumstances. Capturing the “as is” metadata of an organization, its activities, people, and systems, drives a virtuous circle cycle of business mining, modeling, and improvement that provides a clear perspective on how things are operating, and helps shape and define a model of the digital twin.”

A Growing Trend?

With much of the BPO world changed forever by Covid-19, the need for BPO stakeholders to truly know their organization is more pressing that ever. Yet the move to remote work has made processes more opaque.

When the world went home, a lot of companies really found of if they were digital, or if they were ‘and digital’,” said Vaughn. “Lots of BPOs still manage from line of sight. But when that ability falters, and everyone moves home, understanding whether processes are being followed becomes more difficult.”

It’s hard to see why they’re not talked about more in the Nearshore.

AI BPO call center digital twin Manufacturing ML

Peter Appleby

Peter is the Managing Editor of Nearshore Americas. Hailing from Liverpool, UK, he is now based in Mexico City. He has several years’ experience covering the business and energy markets in Mexico and the greater Latin American region. If you’d like to share any tips or story ideas, please reach out to him here.

VIEW ALL POSTS