How to Improve Customer Experience and Slash Operating Costs Using Robotic Process Automation


Share on LinkedIn

By Bill Price, Partner Antuit & President Driva Solutions
and Scott Tweedy, VP Driva Solutions

As business leaders we are all aware that “digital is changing everything”1. As a result, many of us are in the middle of large-scale strategic and digital transformation projects in partnership with our IT teams to achieve the classic win/win: (1) much higher levels of customer experience and (2) lower operating costs.

Even though these efforts are positioned as being agile and customer-focused, they are taking much longer than promised and often fall short of the win/win promise. However, using Robotic Process Automation (RPA) creative business leaders are accelerating customer-impacting improvements without conflicting with the strategic or transformational roadmaps that are in flight.

We will address these three questions: (1) What is RPA?; (2) What can you expect using RPA?; and (3) How can you drive and deploy RPA? Here we will build upon Bill Price’s recent column in CustomerThink “Using AI, Bots, Big Data, and Analytics to Reduce Demand for Support”2 and Scott Tweedy’s recent experience co-leading the RPA initiatives at T-Mobile USA.

1. What is Robotic Process Automation (RPA)?

Robotic Process Automation (RPA) is the application of a technology/software bot, or “robot”, to capture and interpret existing applications for processing a transaction, manipulating data, triggering responses and communicating with other systems. The name can be misleading because there are no actual physical robots doing the work; instead, RPA is business application software that usually sits in the presentation layer and is built to interact on the front-end to digitize data that usually is fraught with manual work or re-work.

RPA is designed to use a software bot to interact with an existing application or website in the same way that a person works with those systems to complete a specific task. RPA not only mimics the actions of a user, but also contains sophisticated business logic to direct the interactions with a system, transform data, and potentially handle exceptions like notifying an employee when human intervention is required. 

RPA tools are designed to mimic the same manual path that are clicked through or navigated by a human, connecting combinations of front-end systems such as agent desktops, user interface (UI), mainframes, or HTML code. RPA tools map or navigate the path taken by an end user in the RPA software for the robot to follow between screens and across various data repositories. These capabilities can be set in motion manually or automatically depending on the rule set that is established to support the business needs including generating critical reporting and tracking capabilities, performing calculations, and triggering downstream activities using machine learning and AI. RPA is initiated and run by the business but deployed in partnership with IT and is fully compliant with governance and security. 

A precursor of RPA is business process management (BPM), defined as “a discipline in operations management that uses various methods to discover, model, analyze, measure, improve, optimize, and automate business processes.”3 However, where BPM attempts to “improve, optimize, and automate”, RPA elevates the art and science to use AI and machine learning to overhaul manual processes with automated tools, the bots, at far lower cost than BPM.

2. What can you expect using RPA?

The business case for RPA is simple, the process to deliver it more challenging so we will cover that in part 3. The business case typically tracks as follows: Reduce manual data entry; eliminate manual data re-entry; and prevent the need for re-work. Much of this is in The Best Service is No Service since it relates to the book’s key principles including “Eliminate dumb contacts” and “Create engaging self-service”4

By mapping the customer journey, you can discover where customer orders or inquiries, or vendor invoices, progress through the organization and where they get stuck, how often they need to be re-entered into different systems, and how long it takes to complete the order or respond to the inquiry or pay the vendor. We are not talking here about averages (see also Bill Price’s earlier column “Get Rid of Average Thinking, Make Every Experience Count”5) but rather a full mapping to show the “long tail”, the longest incidents that of course upset customers or vendors far worse than the “average” order processing time, contact response time, or invoice payment.

Most top-ranked RPA software providers include the following core capabilities:

  1. Process Designer. This is a non-technical code-free graphical user interface (GUI) that allows the business analyst to capture screens, business rules, and workflows. This is the first step as you begin the automation project.
  2. Task Manager. Generally, this capability is provided to add tasks to the automation queue. This tells the software where to go and what to prioritize.
  3. Automation Server. This is the central control system responsible for distributing tasks across the robots and managing workload. 
  4. Performance Dashboard. Provides status and reporting based on business metrics. This capability allows you to see the actual “work” or “output” of the robot.

Netting it out, from RPA you can expect 50% or more reductions for specific and discrete processes in overall process time and associated labor costs. However, RPA can do a lot more than simply reduce labor costs; RPA can also deliver significant increases in customer (and vendor) satisfaction that produces higher levels of customer loyalty and sustainable revenues.

3. How can business leaders drive and deploy RPA solutions?

In order to deliver meaningful change in organizations, business leaders need to transform existing business processes and create differentiated customer and front-line experiences.

RPA starts with selecting the right use cases that need to be aligned with the organization’s strategy for digital transformation as part of an agile automation and modernization effort. On balance it is tempting to try and tackle the biggest, ugliest, and most broken support process impacting your area of responsibility but sometimes a better approach is to start small, prove the capability within a set of business systems, and then gain buy-in across the organization. There are typically three levels of complexity that need to be understood while launching an RPA initiative:

  1. High Complexity. These automation efforts may require deeper IT partnership and engagement (sometimes .net programming), take 6 to 12 weeks to implement, and usually have large long-term cost benefits. 
  2. Medium Complexity. These types of automation efforts could require the robot to query multiple systems and transfer data between systems with decisioning that include “if this – then that” scenarios, and could take 4 to 8 weeks to implement.
  3. Low Complexity: These are automation projects that can be created with the process designer or recorder by a trained or experienced business analyst. These are the repeatable high volume process that we all deal with in applications like Java, Microsoft, or web-based applications, and can be done in 2 to 4 weeks.

Automating these routine processes that require minimal human thought or involvement or re-work enable the organization to focus on more impactful processes and workflows. 

Optimizing the capabilities of RPA is a team effort between the business functions and IT. RPA brings new opportunities to resolve quickly antiquated processes and workarounds often due to legacy system limitations. RPA opens up a new dialogue across the enterprise to deliver immediate value in these three aspects:

  • Eliminating the “swivel chair syndrome”. Frontline reps are not middleware!
  • Stop too many clicks, those processes that take too many system clicks to complete a transaction.
  • Reduce errors, especially caused by manual entry of the same data across multiple systems.

The business teams need to help IT to understand what is in it for them: Automation tools, machine learning, AI, and RPA can optimize business activities and address IT challenges in the age of digital transformation. 

There are many business benefits of RPA

  • Cost Savings and Productivity Gains: Automation of manual processes can result in a reduction of headcount but more often results in freeing up headcount to perform more meaningful tasks.
  • Customer Satisfaction: Automation via RPA can lead to higher satisfaction for both internal and external customers.
  • Business Agility: Can enable faster adoption, acceptance, and execution of agile methodologies. Reduces the impact of inevitable changes to legacy platforms. Supporting unexpected spikes in volumes that are dependent on legacy platforms can be mitigated with RPA.
  • Quality Improvements and Error Reduction: Robots run as they are configured with zero errors and full compliance. RPA reduces errors and enables compliance to regulatory, accounting or security requirements.

How do you know if your organization is ready to explore an RPA implementation?

It is really simple. If your organization has a high volume of repetitious processes or tasks in Java, Microsoft, or web-based (off the shelf or home built) applications, you have a good opportunity to apply RPA as the easiest and fastest to automate these processes, increasing customer experience while slashing operating costs.


1 Title of article in Forbes 11 September 2017, accessed 7 June 2018

2 20 April 2018 accessed 7 June 2018

3 accessed 7 June 2018

4The Best Service is No Service: How to Liberate Your Customers From Customer Service, Keep Them Happy, and Control Costs Bill Price & David Jaffe (Wiley 2008). Based partly on my years as Amazon’s 1st WW VP of Customer Service, but also on “Best Service” providers around the world who have made it easier for their customers to do business with them, we proposed 7 Drivers that start with “Challenge demand for service”:

  1. “Eliminate dumb contacts”
  2. “Create engaging self-service”
  3. “Be proactive”
  4. “Make it really easy to contact your company”
  5. “Own the actions across the company”
  6. “Listen and act”
  7. “Deliver great service experiences”

5 23 December 2015 accessed 7 June 2018


  1. Hi Bill: when seeking candidate applications for RPA, I would look for the simplest transactions involving the greatest volume. Pre-RPA, the earliest memory I have of such a transaction is reporting a missing or damaged daily newspaper. For many years, a customer service representative handled each call, tediously recording the incident parameters along with the caller’s home address. After apologizing, the rep would recommend remediation (taking even more time). And so on.

    Later, the newspaper fully automated the process, giving up the bother of validating whether the customer actually deserved a replacement. The customer’s word was sufficient, and there were only two choices for remediation: re-delivery the same business day, or an extension of the subscription. All automated. As a customer, I view this as a beneficent use of automation. I don’t want to talk to an agent, and the newspaper doesn’t want to incur the expense of supporting my call with a human operator. This works well . . . until my query strays off the ‘happy path’ for the transaction. That is, what happens when my paper is repeatedly mis-delivered? Do I just keep receiving the same transaction path every time I call? Am I offered the option of registering a more urgent complaint? Does software pick up the fact that I’ve called daily for the same reason over the past 10 days? This is where business rules and governance come into play.

    You mentioned governance (“RPA is initiated and run by the business but deployed in partnership with IT and is fully compliant with governance and security.”) but in this context, I think you mean IT governance. When it comes to RPA – and AI in general – the elephant in the room is ethical governance, and it never just happens, and it should never be assumed to exist.

    If my neighborhood is prone to newspaper theft (or high wind?), is my complaint de-escalated because I’m a “frequent complainer” or because I’m otherwise considered low profit to the company? After all, AI can learn this, and algorithms can enforce it. In this case, “it” equals “bias”, which as we know, can result in unfair outcomes. I read today that if you’re African American or Hispanic and want to open a bank account, you will be required to have a larger opening balance than other groups. Algorithms learned that, too. Regrettably, there are copious examples, particularly in the financial world.

    Preventing these types of problems, of course, is the domain of a different type of governance, more nuanced than IT governance. No implementation of AI or RPA should be undertaken without acknowledging the potential for unfairness and bias, and enacting governance – ethical governance – to ensure that incidents are not only identified, but mitigated.

  2. Excellent points here, Andy, thanks. We see “governance” requiring a specialized resource in IT but also the “users”, for example customer service or billing, as well as finance. Totally agree with “start small”. You’re also pointing out some of the fallacies relying on AI, perhaps the next column?

  3. Hi Bill: I think the topic of AI- and analytics-induced bias and unfairness is under-represented in many online communities. The few circumspect voices are smothered under mountains of techno-hype. The best discussion of systemic bias promoted through algorithms comes from Cathy O’Neil’s book Weapons of Math Destruction. Here’s a link to her TED talk:

    The greatest risk I see from AI and robotics is management’s failure to acknowledge the potential for nefarious use, and believing that algorithmic consistency assures fair interpretations and outcomes. I see this hubris regularly. People think that because decision boxes and software aren’t burdened by nuance, they are inherently superior to human judgement. The truth is far different. Along with good results, AI brings unintended negative consequences, and many are gut-wrenching.

    As practitioners, I think it’s myopic to think that technology adoption just provides solutions. We must maintain the additional understanding that with every new solution comes problems – some new, some old. It’s incumbent on every company to explore what they might be, and to respond accordingly.

  4. Completely agree with you, Andy, that AI and analytics could become nefarious. In my 2nd book we cited the Target analytics mess (about a Target customer’s non-pregnancy), and there are other examples coming out every other day. So in addition to anticipating new problems, we need to hew to what my UK-based colleague Peter Massey urges: “You need your customer and business logic to programme a chat bot to do what your customers want, not what you want.”


Please enter your comment!
Please enter your name here