— What processes are best for RPA and how do I prioritize them?

— How does RPA fit in my operating model?

— Which RPA tool is right for me?

— How do we measure intended outcomes?

Successful RPA programs address these questions through five critical steps, starting with an organizational macro assessment of current processes (along with their tactical and strategic initiatives) and the technology landscape affected by RPA.

Related: Insurance 2017: Priorities for innovation, automation and transformation

Step 1: Perform a macro assessment.

A comprehensive macro assessment of processes that support your lines of business (LOBs) is a critical step for optimizing and automating processes using RPA tools. This macro assessment will lead to building a roadmap for you company's automation journey that is aligned with the company's vision and goals, will maximize ROI and lead to a successful implementation of the RPA solution.

Given the complexity and uniqueness of today's insurance carriers, it would be naïve to think that a one-size-fits-all approach to RPA would work across the industry. Therefore, examining your organization from the top down by functional business case is vital to determining what type of RPA solution will fit best and how it should be deployed.

Take, for example, the claims process: create a top-down view by breaking the end-to-end process into functional groups (intake, policy verification, segmentation, fraud verification, adjudication and payment approval) and analyze the number of resources aligned. Then identify the individual processes that make up the functional group along with the respective headcount and volumes.

Related: Realities and risks of the robot revolution

While the macro assessment is underway, you can begin to define an operating model, perform process assessments and select a tool or tools.

The following Gantt chart illustrates how you can plan to run various activities while catering to dependencies. For example, "Implementation" cannot start until a tool has been selected.

Some insurance industry leaders have built Enterprise Centers of Excellence to provide oversight on new projects and ensure that they are aligned with the organization's strategic automation goals. (Photo: iStock)

Step 2: Define your operating model.

To gain competitive advantage in the marketplace, RPA requires rapid implementation.

Yes, a central enterprise group should set the required governance and controls to ensure consistent standards. But organizations that follow a centralized model for all things RPA will lose their speed to market, anticipated efficiencies and competitive advantage.

Across the insurance industry, many organizations view RPA as any other technology that demands control rigor from within IT (heavy SDLC, all RPA configurations centralized into IT, etc.) This has led to multiple cases where initial RPA efforts have been mired with missed deadlines, cost overruns and ultimately failure to achieve the initially identified benefits case. 

Conversely, the most successful leaders in the industry have chosen to centralize governance, standards and frameworks, but choose a tool (see Select the Right Tools) that allows them to configure the Bots within the business or business technology group and teach it to high-end business/systems analysts in a matter of weeks (or a few months at most). Additionally, leaders in the industry have built Enterprise Centers of Excellence that are engaged with the individual LOBs and their intake models. In doing so, the CoE provides oversight on new projects, ensures that they are aligned with the organization's strategic automation goals, and that they are leveraging existing tools.

As part of the governance, the COE also plays a central role in establishing and monitoring key performance measures.

See also: 6 steps to take to evaluate cyber risk

Too many organizations view RPA assessments as a one-time activity, which can be a mistake. (Illustration: iStock)

Step 3: Assess processes.

Most organizations do a fantastic job identifying processes for RPA. For example, processes that were previously BPO'd and processes that are highly repetitive with heavy staffing required to support them.

Yet, too many view this as a one-time activity and fail to continue the analysis across all LOBs. Simply identifying processes as a one-time activity and moving toward design and implementation leads to extended timelines and costly technology workarounds to bridge RPA gaps.

If struggling to identify candidate processes, take the processes that make up the functional groups identified in the macro assessment and assess them from a bottom-up perspective. Continuing with the claims example, for the payments process, assess each process by reviewing all relevant SOPs, process flows and user guides, and determine which applications are used throughout the process. You will ultimately determine the applicability of the process from an RPA perspective based on two things:

1. The data necessary to complete the process is discrete and digital.

2. All business rules required to run the process are clearly defined.

Once you determine how much of the process can be automated, you can estimate the expected benefits and build the business case specific to that process. The business case can be aligned to revenue creation, process optimization and/or cost reduction or consistency of meeting regulatory requirements, but regardless of the metric, the business case should be clearly defined prior to proceeding with the automation project.

Related: 5 high-tech challenges (and solutions) for today's independent agents

Make process assessment an ongoing activity to continuously find opportunities for cost savings within the business. Establish an intake process capable of continuously prioritizing the backlog of RPA projects for delivery.

If selecting the first RPA tool for the organization, you should focus on a tool that will enable you to achieve rapid results based on the requirements within your organization. (Photo: iStock)

Step 4: Select the right tools.

A macro assessment will enable your due diligence process prior to signing with an RPA tool vendor, so that you make the best initial decision. Each of the tools on the market has strengths and differentiators that make them the valid choice for various firms. We have seen multiple examples where a client has selected an RPA tool prior to completing a true analysis, which has led to committing resources (time, labor and money) to a tool that does not enable deliveries that align to the organization's ROI or timeline expectations.

Essential questions a macro assessment will help you answer before selecting a tool include:

— What kind of automations will be executed: assisted or unassisted?

— What data ingestion capabilities exist, and what are needed: OCR and/or ICR? 

— What capabilities do the business rules require: basic RPA or cognitive? 

— Where will the automations occur: business processes or IT automations? 

— Who will be configuring the bots: technologist in a centralized model or business analyst in a federated model?

If selecting the first RPA tool for the organization, you should focus on a tool that will enable you to achieve rapid results based on the requirements within your organization.

Related: The impact of more data on the claims handling process

However, if you already have a tool and realize another option may be better suited, be prepared to leverage multiple tools depending on the complexity of the process being automated.

As you begin to scale your automation efforts, you will encounter a unique set of challenges for which you will need to prepare. (Photo: iStock)

Step 5: Implement in modular fashion.

Take the time to do a bottom-up view of the process you've selected for RPA. By doing so, the organization will likely identify several key steps where deployments can occur using a modular, or minimum viable product "MVP," approach so benefits can be realized rapidly. Understand that RPA is not the answer to every need, and other automation technologies (e.g., data ingestion and cognitive) may be required to achieve certain automations or to enable Straight Through Processing "STP."

Start by going after the low-hanging fruit such as the "swivel-chair functions" within the claims process with the highest applicability for RPA that can be delivered rapidly at a low cost/risk. Build an inventory of those functions, such as claim entry, processing and payment across the applicable processes. Then, stack-rank them based on immediacy and scale of payback. Create an RPA intake model that will soon provide a self-funding mechanism to pay for further RPA efforts, as well as the deeper automation technologies required for the more complex scenarios.

Related: Here's how auto technology will change claims

By thoughtfully planning your approach, you will set yourselves up for success.

Paul Krauss is FSI Technology Practice Lead at NTT DATA Services. Dilip Kumar is FSI Director at NTT DATA Services. John Petrusick is FSI Principal Consultant at NTT DATA Services. To learn more, visit nttdata.com.