In most enterprises, automation can easily provide value in the spaces between a business function’s IT systems and the company’s business processes. These spaces include people: the employees or contractors who perform the repetitive work required to process data and shuttle information from one system to another. Legacy systems that do not work well with newer technologies are another common space where automation can be applied to improve systems and processes.
But IPA is not a quick-fix, set-it-and-forget-it solution. It is, rather, a significant initiative that requires thoughtful planning, application and expertise to ensure optimal return. Companies must develop a keen understanding of how automation will be applied to the current and future state of their processes and systems. Even in ideal circumstances, an IPA initiative requires careful oversight. And if it is applied atop broken processes or unstable systems, companies will quickly find themselves in a continual change management cycle, jeopardizing stability and value to the business.
Here are six obstacles that can slow or derail a successful IPA project. (In part two of this series, we’ll explain how to address these impediments.)
In a typical IPA project, subject matter experts (SMEs) — those currently performing the work to be automated — are required to explain the work they perform to support each bot’s design. SMEs might resist fully explaining their tasks, as they fear that their skills might not be needed after training their non-human replacements.
Before engaging these SMEs’ insight, organizations must plan to address the people side of automation. How will the company express the value of their input and explain what lies ahead for these workers? Will they be retrained or move to a different, more engaging role? Will they be asked to continually revise requirements and perform work manually when new scenarios or edge cases are encountered? Without the right upfront engagement, SMEs may not be as cooperative as needed.
When companies fail to keep documentation up to date — a common occurrence even in strong organizations — the risk increases that critical steps will be missed in the planning and implementation of automation. Dubbed “hidden factories,” these undocumented routines and workarounds tend to crop up and multiply when a standard operating procedure becomes outdated. These types of additive actions require far greater SME input and programming to ensure that they are included in the automation plan.
Consider, for example, a multinational bank with thousands of employees who perform manual work largely dictated by the industry’s ever-changing regulatory environment. The need to keeping pace with change while addressing cost pressures might leave little time for staff to spend formally updating and documenting numerous manual processes.
Even well-documented processes become fragmented over time. When a process or task is handled by different workers, it grows progressively more reliant on human nuances rather than rules and procedures. Individuals might develop personal cheat sheets and rules of thumb to complete tasks, and these shortcuts often don’t make their way into the automated process. That’s a problem because the success of the automation is absolutely dependent on the process being as close to complete as possible 100% of the time. Moreover, operational processes tend to grow out of immediate work demands. As a result, they are not thoughtfully designed to maximize throughput and minimize human intervention — that is, outcomes that a human must handle because the steps that process those items are inaccurately captured or missing from the process automation altogether.
The health insurance claims management process presents a good example. A bot reviews each claim to verify the accuracy of multiple data points, then checks the claim against the payer’s payment policies. When the bot is built, every possible outcome that doesn’t require manual review by an employee must be accounted for in coded instructions; otherwise, the bot cannot successfully process the claim, resulting in the need for human intervention.
Due to the growth of IPA, automation specialists and engineers are in high demand, as is the need to create a center of excellence (CoE). In “RPA Operating Models Should Be Light and Federated,” Forrester Research estimated that for every 100 full-time employees whose tasks are targeted for automation, 10 to 15 new roles are needed to manage the initiative.
This group of managers, made up of stakeholders, business analysts and engineers, becomes the automation CoE. This CoE plans, designs, implements, manages and maintains the initiative for the life of the program. The chosen IPA platform also dictates the level of expertise required to develop and apply the initiative to the company’s specific environment.
Some IPA platform vendors might make the implementation process appear easier than it actually is, citing a process discovery solution and pre-built bots. They might also say that virtually any CoE member can build a bot, which isn’t entirely true. For business applications with little to no customization, pre-built bots can be deployed with relative ease. It’s quite rare for a business to implement an application straight out of the box, with no customization. Taking automation beyond baseline functionality requires skilled engineers.
According to Acumen Research and Consulting, the global IPA market will hit $4.1 billion by 2026, with a 32% compound annual growth rate between now and then. Contributing to this growth is the overall increased interest in digital initiatives. Organizations large and small are looking for ways to create synergies across their business processes and technologies, and have recognized IPA as a technology that can aid in their execution, at scale.
Heeding the hype, many companies jump into automation without proper planning and analysis, much less an overall strategy. In a Forrester study, “Gauge Your RPA Maturity,” less than 1% of companies achieved multiple bot deployments across all business segments with a formalized operating model. In addition, robotic process automation (RPA) pricing models are notoriously complex, making it even more difficult for organizations to calculate the total cost of ownership (TCO) and return on investment (ROI).
HFS Research predicts that licensing costs represent just 25% to 30% of the total program cost, with 65% to 70% attributed to the cost of supporting the program. Moreover, there is a significant ongoing cost to address maintenance. Uncertainty around what to automate and what it will cost the business compared to what it will deliver prevents organizations from realizing the full value of IPA.
An IPA program requires coordination across the organization to automate the right processes and establish guardrails. In the absence of this coordination, the downstream impact is increased friction, new process gaps, and increased operating cost.
Bots, by nature, break and require maintenance as business processes, and the applications they utilize, evolve. Bots may also step on the work of other bots, causing a cascading effect of failed outcomes. Security and compliance are additional factors that must be governed. What access credentials do the bots need? Which bot performed what work, and when? What went wrong in this particular process, and why? Who is responsible for fixing this? And how do we calculate the business cost incurred because of these missteps?
An IPA initiative that is misaligned, mismanaged, siloed, or unmaintained cannot scale or deliver value to the business.
IPA programs have outstanding potential, and businesses are already realizing the benefits. However, the impediments discussed here all stand in the way of success. In the conclusion of this series, we’ll explain how to surmount these impediments.
Learn more about how to avoid these common pitfalls in part two of this series, “How to Obliterate Obstacles that Hinder Intelligent Process Automation,” and by visiting the Digital Operations section of our website, or contact us.