Bryan Bearden
Bryan Bearden

Many companies are beginning to report a substantial return on their investments in process automation, reaping benefits like financial synergies, redeployed human capital, and improved speed of delivery and quality. Yet despite these well-documented success stories, a large number of enterprises are still struggling to gain traction on their journey toward the meaningful exploitation of automation technologies. Why? While these programs often begin with superb executive presentations and generate initial optimism, execution stalls due to competing investment priorities, resistance from non-supportive stakeholders, technology integration challenges, and over-extended IT and business resources.

The pandemic has changed all that. For leaders, companies, and entire industries, COVID-19 has served as a “wake-up call,” igniting interest in strategies and ideas they had long deprioritized. Business leaders cannot afford to face the future without reimagining how work gets done. Facing that new reality means, for example, a newfound reliance on videoconferencing and workforce geographic diversity. It also entails a significant exploration of ways to leverage automation technologies to provide insurance against workforce-impacting events and to enable distributed team members to function effectively.

The challenges are notable for some industries. In healthcare, for example, unprecedented levels of complexity combined with palpable misalignment between payers, providers, and patients create daunting transformational headwinds. But that industry, in particular, cannot ignore the upside opportunities available by leveraging technology to reduce cost, improve patient outcomes, and develop new health-impacting capabilities. While other industry verticals have also been slow to embrace opportunities for the automation of human work and human thinking, that is fortunately starting to change. In this blog, we will describe the steps traditionally slow adopters, such as healthcare entities, can take to gain meaningful momentum in the adoption and use of automation technologies that can drive improved business outcomes.

Automation Technologies: Defining Terms

When talking about automation, many commentators use the generic term “AI/ML” to describe a variety of software applications and technologies that mimic aspects of human intelligence. At Wavestone, we prefer the more inclusive and descriptive “automation technologies” to reference the broad spectrum of technical capabilities that support enhanced human/machine interaction.  As we present the various components below, notice how they build upon each other and work together to produce high levels of automation and business value.

  • Robotic Process Automation (RPA): RPA is the use of technology to automate repeatable, consistent tasks that humans usually do on a computer. These tasks can be as simple as copying data from a spreadsheet into a data processing system or automatically routing transactions to human processors based on a set of rules. The distinction here is that the “machine” in question is not making judgment calls; instead, it follows a very repeatable set of requirements and rules. When exceptions are encountered, a human is required to resolve the anomalies.
  • Artificial Intelligence (AI): While RPA can be considered a juvenile version of AI, we think of AI as a progression up the human-thinking maturity curve. AI moves beyond systems programmed to follow basic rules into the world of decision making, replacing what might often be considered a judgment call. AI is a broad concept that leverages capabilities like machine learning and predictive analytics.
  • Cognitive Automation: Cognitive automation represents the application/execution of AI concepts to automate human decision-making. This approach could be as simple as using intelligent OCR to scan and interpret documents, replacing data entry processors. Or, in the case of the healthcare industry, it could be as advanced as using machine learning and predictive analytics to “score” claims with the likelihood they will pay incorrectly and result in overpayment. In many industries, it could even replace portions of rules-based determination/routing logic and processes.
  • Machine Learning (ML): Machine learning is the process of training “models” (advanced algorithms) to recognize patterns in data. Continuing with a healthcare example, we can see how ML would work. By ingesting and comparing two years’ worth of medical claims history, a model can “learn” to identify correlations between unrelated data sets. These results could allow a cognitive automation engine to identify which claims are likely to result in problems based on historical patterns in the data.
  • Predictive Analytics: This concept uses the results of ML to anticipate the future. For example, when an ML model scores a particular intake submission as “highly likely to result in rework,” it predicts an outcome that can inform action on the part of a system or operations processor. For example, predictive analytics would route claims identified as likely to be problematic to human experts for processing vs. others that the system would send to an RPA solution to complete appropriate actions.

Wavestone’s Take on Automation: Start Fundamental, Imagine Monumental!

While each company’s situation is unique and will require custom analysis and architecture, the following suggestions, taken in order, will help you as IT and business leaders move forward with your automation technologies programs:

Getting Started
  1. Document, analyze, and rationalize current business processes.
  2. Identify redundant activities that can produce early quick-win automation value (low-hanging fruit).
  3. Identify areas where a virtual robot can automate work to free up human capital to do more strategic tasks and reduce overall costs.
  4. Implement simple RPA scripts, measure results, and expand.

Key considerations during this phase:

RPA Packaged Software

  • Several software vendors have created easy-to-use and fairly robust RPA platforms that clients can use to create automation scripts.
  • Top consulting firms have effectively used these packages to help companies understand and develop solutions to kick-start their clients’ automation technologies programs.
  • A word of caution: While the marketing from RPA software providers and consulting firms can be compelling and the configuration of RPA tools is quite intuitive, we have found that many business users are not “wired” to think in terms of end-to-end considerations and non-functional requirements (like systems performance and efficiency). The burden inevitably falls back on IT. Which is why we recommend…

Automation Center of Excellence (COE)

  • Implement an Automation Center of Excellence (COE) to provide governance, standardization, and integration of various RPA and automation technologies solutions.
  • The COE will oversee in-house and third-party automation efforts to ensure RPA scripts perform properly, interoperate with various applications, data stores, and hard-copy forms, and are maintained/enhanced to achieve the expected levels of quality and efficiency.
  • By setting and enforcing standards, the COE promotes the creation of automation assets that various business units can adapt and reuse.

Moving Ahead

  1. Move towards AI/ML/cognitive automation by identifying small candidate areas to pilot (e.g., vendor demographic clean-up effort, paper submission intake, customer service chat-bot, etc.).
  2. Deliver high-profile minimum viable product (MVP) projects as small showcase wins, and understand your company’s appetite to grow in that area.
  3. Consider hiring a data scientist and begin exploring opportunities for machine learning.
  4. Use successful RPA implementations to introduce higher levels of automation technologies, such as cognitive automation and predictive analytics.

Key considerations during this phase:

Automation/AI Team

  • Regardless of any vendor or integration promises, automation technologies cannot be effectively implemented without heavy business engagement.
  • We therefore recommend that you build teams dedicated to automation that are a hybrid between each business unit and IT. These teams act as “spokes” to the COE hub:
    • Business unit members identify candidate functions and processes for value improvement, develop requirements, establish funding for the business case, and champion the change management effort within their organizations.
    • IT resources perform technology/package selection, provide data scientists, and perform coding, configuration, and systems integration and maintenance.

Building Business Cases

  • Build realistic business cases. Don’t fall for the “best-case scenario” promises from vendors and integration partners. They may have statistics (even data from your own systems) behind the pitch, but they rarely consider broader data and process complexities. It will take time to realize those “ideal” numbers.
  • While automating something is easy, automating the right thing at a volume that makes a difference requires hard work, patience, and trial and error. Temper expectations to avoid having an initial false start derail the entire program.

Imagining More

If you get here, congratulations on having developed a comprehensive approach to automation technologies! These capabilities will continue to mature, so your journey will need to evolve as well. You can add your own imaginative ideas to our short list:

  1. Imagine “smart edits” that use pattern recognition to identify invoices/claims/other intakes that do not match the typical payment “pattern,” replacing the hundreds (or thousands) of edits. Even more imagination: a data scientist who constantly adjusts and “teaches” the model and replaces dozens of human processors?
  2. Envision a set of “self-healing” financial transactions that keep balances constantly in sync, combining ML-based AI to identify and RPA to correct.
  3. How about an algorithm-based “robot” that analyzes process maps, user activity, and other data to find and document additional automation opportunities?

As with anything that makes a difference, implementing automation technologies is more of a journey than a destination. This journey can seem daunting, especially if you haven’t started to leverage the potential of these powerful tools. Wavestone’s advice is to start small, build incremental successes, and keep evolving your vision of how these capabilities can build on each other to provide exponential improvements in business performance.

Rest assured that the results are worth the effort. The hardest part is just getting started!

Bryan Bearden
Lead Principal

Bryan Bearden is a senior transformational IT leader, advisor, and strategist, with over 20 years’ experience supporting healthcare, manufacturing and software services. Specialized in application development, cultural modernization, and automation, his career is highlighted by a host of transformational software and system projects ranging from blockchain solutions to mainframe consolidation and cognitive automation.

The Case for Contract Lifecycle Management in Your VMO

Jul 22, 2021

The implementation of a mature contract lifecycle management (CLM) program allows for continuous visibility of documents and reduces the loss of their control.

Mergers & Acquisitions: Key Considerations for Technology Integration

Jul 20, 2021

There are several elements that should be considered when planning for a technology integration following a merger or acquisition.

Have a Question? Just Ask


Whether you're looking for practical advice or just plain curious, our experienced principals are here to help. Check back weekly as we publish the most interesting questions and answers right here.

Ask Wavestone