How robotic process automation and cognitive technologies will transform the insurance industry
A strategy centered on the needs of today’s consumer
Pilot programs exploring robotic process automation (RPA) in insurance are showing positive results and benefits that go far beyond efficiency gains. The potential for robotic and cognitive automation across the insurance value chain is significant.
Robotics and cognitive technologies
While RPA is a first step, automation is expected to be increasingly driven by cognitive technologies in insurance. Here we explore how insurers can reconfigure operating models and adopt a more customer-centric approach to capitalize on the opportunities unlocked through cognitive technologies and robotic process automation.
Using robots to drive tangible business benefits is very much a reality today: The IT-enabled RPA market has been growing rapidly at a CAGR of 60.5% from 2014 and is expected to reach US$5 billion by 2020.1
RPA in its pure form, however, is just the beginning. Cognitive capabilities that enable machines to perform tasks normally reserved for human intelligence are now being leveraged with robotics as well. Cognitive technologies include such capabilities as machine learning, natural language processing (NLP), machine vision, emotion recognition, and optical character recognition, among others. Each of these technologies builds on the existing competencies of RPA and advanced analytics, including neural networks, data mining, and Big Data processing.
The resulting combination—termed robotics and cognitive automation (R&CA)—encompasses a potent mix of automated skills with application across the insurance value chain. R&CA is expected to foster greater collaboration between human and machine by both automating repetitive tasks and enhancing the quality of jobs. R&CA technology is now poised to unlock a world of possibilities through the synergistic combination of its key components.
Impacts to the insurance operating model: People
While adoption of any new transformational technology warrants reconfiguration of the insurance operating model, cognitive technologies and robotic process automation in insurance are likely to have the highest impact on the people and technology aspects of the current operating model.
In most insurance organizations, the current delivery pyramid is significantly bottom heavy with the majority of volume-heavy transactions and reporting processes (e.g., regulatory reporting, claims processing, document verification) being performed by humans. Robotic process automation in insurance will likely reshape this pyramid as insurers automate many of these transactions/processes, potentially reducing the size and engagement of the bottom and middle layers of the delivery pyramid, with growth in the top layer.
- Business development, product, and marketing jobs will increase due to demand for skills in areas such as data analytics, machine learning, and development of algorithms.
- Operations, including policy servicing and reporting, will have ever-greater levels of self-service and automation, as well as completely new, highly streamlined digital processes.
- IT and other support functions will require fewer FTEs due to reduced overhead through standardized and automated processes, and the potential migration to new strategic data platforms on cloud and other third-party analytical tools.
Impacts to the insurance operating model: Technology
Insurers should prepare themselves for the imminent R&CA transformation by reconfiguring their IT systems. The transformation will be an extension of the journey that has begun in such areas as RPA and advanced analytics enablement, most likely including:
- Modular sourcing: The R&CA technology industry is now engaged in a startup-like phase, in which nimble firms that provide specialized technological capabilities are well positioned to disrupt the incumbents. These vendors are already providing disaggregated services on the cloud. For example, one leading ecosystem player provides a series of modular services such as "personality insights," "visual recognition," "text analytics," etc. Through this approach, insurers can source different capabilities from niche vendors provided that their underlying IT architectures are flexible enough to tap into cloud-based services and use them as "cognitive operating systems" in building intelligent applications.
- Integrated systems: R&CA technology has the inherent capability to iteratively self-learn and generate insights through access to data from multiple sources. To maximize the returns from this technology, integration with legacy systems and other emerging technologies such as Big Data, IoT, and cloud must be achieved. This will likely lead to a much desired breakdown in silos of data across the enterprise, enabling the establishment of a single source of truth and delivery of a unified data model in a significantly more consumable form.
- Transparency and control: Cognitive technologies and systems will undoubtedly partner with humans in the near future. To gain trust in the robots, humans will need to understand how a particular decision has been reached. Given the nascent stage of R&CA technological development, humans are expected to have the ability to overturn machine-made decisions. Furthermore, regulators are likely to insist on robust audit mechanisms. The R&CA systems of the future will have to be designed keeping in view all these transparency and control features.
The technological landscape is evolving quickly, with two crucial implications for insurers: 1) the need to identify and source relevant capabilities to allow for better task design; and 2) an appropriate division of labor between humans and machines.
How should the industry approach an R&CA-driven transformation across its value chain?
To start, the core foundation of R&CA technology is guided by four broad principles:
- Cognitive systems need to synthesize vast amounts of data to generate powerful insights and connections.
- Robotics capabilities play a dominant role in performing actions that would otherwise be driven by humans, the results of which produce insights that help generate and validate hypotheses to aid in decision making.
- Cognitive systems must interact with customers and employees using natural language and demonstrate contextual reasoning.
- Given their probabilistic nature, cognitive systems need to continuously learn from their past actions and evolve more accurate algorithms.
With a firmly established core foundation, insurers can adopt a measured approach to achieve their desired R&CA goals.