The RPA Progressive Journey

Some consider Robotic Process Automation (RPA) a fad. They see it as an interim solution to stitch together processes that have long been ignored and eat up time and resources, but they believe they will one day re-engineer or eliminate as part of a larger transformational strategy. I believe those people don’t see the bigger picture.

Sure, transformational strategies will change how existing processes are performed and the level of automation within those processes. But RPA is more than just temporary glue to cross that chasm. It’s the software that builds the bridge from the as-is to the to-be and then facilitates the long-term vision of automated end-to-end processing. The trick is getting businesses to understand how RPA can not only support the transformation but enable continuous improvements.

I like to use the following as an example of a journey that customers can use to target RPA’s role in process transformation.

Phase 1: Emulate human in front of the computer

Phase 1 is where we find many RPA customers today as it is currently the most common use case. Humans are spending significant amounts of time on tedious low-value tasks. Out of the box, RPA can model the human’s activities through the ability to read application data off the screen, emulate keystrokes and mouse movements, and implement simple decision-making. The outcome from this ability alone has been tied to significant upside benefits such as improved customer service, increased speed to completion, thousands of man-hours redirected to higher-value tasks, increased quality of data, and improved employee job satisfaction.

Yet, it’s far from the end of the story and the value that RPA can continue to bring if businesses continue along their journey.

Phase 2: Extend emulation to include analysis

Phase 2 offers tremendous opportunity in the age of democratized cognitive solutions. This is where we can peek into the processes that were previously handled by humans and start to incorporate intelligence from those tasks that were most likely not previously captured. Those include complaint emails responded to daily, the most often mentioned product in customer communications, and the amount of money associated with discrepancies between what was billed and what was received.

It is possible to also see this data in some systems of record, but often that requires analysts to actively plan and coordinate the value of that data, implement the reporting surrounding that data, possibly integrate data across multiple systems, and then analyze that data. Yet, RPA can simply capture the data as part of task execution and enable users to explore and correlate the data quickly and easily. I believe it’s one of the most underutilized aspects of RPA programs and one that offers tremendous insights into the workings of the business supporting overall better decision-making. I’d go so far as to say that data captured during this phase may never even make it on the radar of business executives, sacrificing potential jewels.

Phase 3: Extend automation to human-assisted decision-making

The next phase in the journey is to automate human-in-loop tasks. These are tasks that require human intervention such as exception handling or approvals. There are some businesses that start with this use case as their goal but, ultimately, need to still be successful with phase 1 as those are the automated tasks driving the data collection, encountering the exception, or orchestrating the need for approval.

As businesses progress along their journey, certain processes will not proceed beyond this point, and that’s okay. For whatever reason, the business may require that a human needs to be an actor within the process; however, some processes will eventually move on to phase 4.

Phase 4: Extend automation to machine-directed decision-making

In phase 4, we leverage all the work in phases 1-3. We have automated the capture and handling of data at the start of the process. We have analyzed and looked for patterns within the data that is moving through the process. For some period of time, we have trained an intelligence layer by having a human participate in the process. And, now, we have reached a point where artificial intelligence (AI) has enough data and training to arrive at an acceptable process conclusion without human intervention. Moreover, since most AI operates by computing probabilities, in cases where the business determines that the AI “score” is not acceptable, we can fall back to a phase 3 approach to allow a human to review and make a final decision.

As you can see, by setting our sights on arriving at phase 3 or 4 as part of our RPA strategy, RPA should occupy a much bigger role in transformational efforts. It’s easy for the tactical needs of business, such as those that match automations aligned with phase 1, to overshadow a more comprehensive long-term vision. That is, because RPA-based automation delivers measurable value when applied in the simplest cases, it may inhibit activities to explore more strategic use cases. In addition, and quite often, those making the choice to use RPA to satisfy phase 1 requirements are not often the same individuals responsible for looking at how RPA corresponds to a greater transformational effort.

In the long run, businesses that actively participate in defining and guiding the RPA journey will see great returns and reap the benefits of its tactical and strategic capabilities.