This latest Opinion piece is by Aman Rangrass, the Global Head of Revenue and Partnerships at Skan.ai. He has long been a student of understanding how work happens with customers in the insurance and financial services verticals. Aman has more than 15 years of experience, with a focus on large-scale transformations from opportunity identification through implementation as part of their business transformation journeys at McKinsey & Company.
Sometime in 2018, automation officially caught fire. It became a core focus of nearly every organization’s technology and operations strategy. At the time, an unacknowledged struggle was also quietly playing out. The headline of this struggle was a high failure rate and an inability to scale automation at the enterprise level. Ernst & Young would later reveal that 30% to 50% of automation projects fail. These headlines were only symptomatic of a deeper malaise: that enterprises were driving blind on how they work.
Today, automation is still top of mind, but the narrative has shifted. The class of problems insurers are looking to solve has evolved. Market conditions are motivating executives, out of necessity, to dig into that deeper understanding of how work gets done. Every enterprise, whether implicitly or explicitly, is looking to do “more with less,” and with a renewed sense of urgency. Talk to senior stakeholders at any insurer, and it becomes obvious they are looking to marry their existing technology stack, the operators central to business-critical processes, and the very processes themselves. While much of this is centered around the claims function, the chasm of inefficiency and ineffectiveness is seen everywhere and everyday.
The fundamental issue at play, on which stakeholders universally agree, is a lack of understanding of how work gets done around three axes – cost, quality and personnel. For example, what is a granular view of the end-to-end claims process? What are the different permutations, exceptions, activities and costs of the variance between the observed versus the as-designed process? What are the drivers of quality or missed SLAs?
In response, most insurers will deploy a team of business analysts to run interviews. However, this does not provide us a true picture. Hybrid solutions, often referred to as process intelligence, are now emerging to ensure greater depth of accuracy for process improvement planning, workforce efficiencies, and digital transformation.
Laying the Foundation for an End-to-End View of Work
There are four major pain points currently plaguing insurers. First, insurers don’t have a granular breakdown of the cost associated with different levels of service for end customers. While this has usually been done through manual observation and time tracking, there is less color in the complexity associated with servicing the more complex customers and cases. Lastly, time to proficiency for new operators is tough to gauge. Most insurers assume a linear ramp-up based on historical data despite knowing that industry context and knowledge matters and their processes continue to evolve over time. Furthermore, hiring a younger workforce is challenging given the optionality from other employers. This requires raising pay rates and finding new ways of incentivizing work.
Next, insurers have significant technical debt from the adoption of multiple home-grown and third-party technologies, making it difficult to understand the use and impact of all the different applications employed as part of their daily operations. Third, gaining transparency into employee utilization across the entire workday is difficult, especially in an environment where there is a mix of onsite and remote employees. For example, operators in some processes spend 10% to 20% of their time self-reporting activities, which takes away from daily operations and is generally overestimated.
Lastly, knowledge and workforce retention continue to be a pressing issue. While the exact figure varies, many insurers claim that more than 50% of claims organization personnel will be “eligible” for retirement by 2025, which suggests that significant institutional knowledge of core business processes and decision making will be lost.
Tackling these pain points starts with a nuanced understanding of how work gets done. Many insurers use some form of manual time tracking to generate an understanding of their processes, and some have turned to approaches such as process mining to build a more granular understanding. However, the claims process, for example, does not yield itself well to process mining as operators work across multiple applications to complete a case.
Manual tasks also make it difficult to capture how work moves from one participant to the next and the associated wait time between steps. Additionally, tasks in one process may have a dependency on an upstream process and that relationship is lost if the underlying data is either narrowly defined or sits on different application backends that require significant effort to link the data.
For many insurers, processes are executed based on an individual worker’s knowledge of how to complete the task with limited documentation about the workflow, participants, and manual redundancies. Without a basis for understanding these nuances, customers will only have visibility into some cross-section of work if they do this observation manually or by looking at only a single application. Developing a complete end-to-end picture is necessary to understand process improvement efforts and the ability to assess potential automation opportunities. Importantly, doing this quickly, easily, and accurately is essential when every enterprise is under pressure to reduce costs.
Customers are now acknowledging this non-negotiable step: conducting rapid data-driven process discovery with the right sample size of users in order to create a true end-to-end representation of work. This provides a holistic view and highlights opportunities for process improvement and standardization, before undertaking automation or other interventions to unlock value. Working through these steps sequentially will impose the right organizational rigor and yield efficiency gains that are more sustainable. The data-driven approach can then be used to measure and track improvements and results over time.
Unlocking the Value Trapped in Your Processes By Disrupting the Status Quo
A new hybrid approach of “process intelligence” solutions are emerging to address these gaps by extending the use of artificial intelligence (AI) to provide an end-to-end view with greater depth of accuracy for process improvement planning, identifying opportunities to reduce error and rework, workforce productivity, and automation.
An insurer, challenged with finding a double-digit reduction in operating costs, was looking to develop an understanding of the manual steps in their processes with a view towards implementing automation. A large cross-section of agents executing two distinct processes were monitored over multiple months to gather a sizable representative dataset. The first process was highly variable with no user following a standardized path. As a result, any automation implementation would only have added to their technical debt in this process. Instead, they focused on process standardization before being automated. In contrast, the second process revealed that greater than 60% of the cases followed a standard path, and automation was implemented on select steps to help optimize the process. A similar approach is being employed enterprise wide.
Another insurer was looking to enhance their operational and financial performance measurement, through a KPI-driven view of employee utilization, productivity, application usage and, ultimately, labor cost associated with processing an end-to-end case. The insurer made data-driven decisions on current and forecast workforce productivity based on an understanding of staffing based on operator effort and not case inventory. Forecast operator hiring plans were reduced by 25% and existing operators reduced their manual tracking of hours. Additionally, insights into the use of the different applications used to execute an E2E case, including the number of actions on each application, allowed the insurer to prioritize product development opportunities based on the highest pain points for operations.
While such approaches that employ AI and scalable data are not silver bullets to unlock value trapped in processes such as end-to-end claims management, a data-driven approach will yield new insights that will inform sustainable long-term decision making.
Be the first to comment