Elaine Mannix is an insurance leader at UiPath.
PwC states in their 2019 Insurance CEO survey that 70% of today’s global insurance CEOs want operational efficiencies to drive growth. One of their top challenges is system integration. System integration challenges have been around for what seems like forever in the insurance industry. Never is this more apparent to a policyholder than on their claims journey. And never has this fact impeded the insurance industry’s ability to drive operational efficiencies more than today due to competitive pressures.
Insurance digital transformation opportunities make my job at UiPath—and being part of the Robotic Process Automation (RPA) community—so invigorating. As an insurance industry veteran, I see new automation capabilities as just-in-time propositions for those who are looking for efficiencies to drive revenue.
The technology is finally here to break through the industry’s legacy constraints without the usual hefty price tag. RPA is a significant change enabler! And I get to be a part of helping customers see the potential of RPA, plan and implement massive change, and think ‘automation first.’
But how are insurance leaders fairing right now with their operational efficiency goals? This blog post takes a look by way of my ‘boots on the ground’ conversations about automated claims processing. So far, efficiency gains are piecemeal and don’t meet the expected potential, particularly when applied to legacy systems or back-end quick win processes. Their starting points were usually burning issues or ‘low-hanging fruit.’
Many insurance providers have good lessons learned, but these relate to siloed quick wins only. These quick win processes tend to be related to moving information in and out of legacy and backend systems. This type of RPA project, while essential, won’t move the needle on profits.
A lot of people think about insurance companies as stalwarts, or vulnerable to displacement by insurtech, but this isn’t the case. I see the incumbents’ potential every day. They have the skills, experience, and data to drive growth with new, dynamic products and services. The incumbents also have the original ‘data forerunners’ from data scientists to actuaries. What’s needed now is the capability to unlock data, gain new insights, and get it to the right people at the right time.
My take is that insurance industry disruption will come from within, where key talent is unhappy with the pace of change. This talent are already protagonists who are eager to take on new opportunities.
The insurance leaders who are moving ahead of the pack right now are taking a holistic view of automation initiatives to drive value. They’re doing this by prioritizing these activities:
Automating ‘across the top of the organization’ (this means aligning work to key strategic growth initiatives for growth, efficiency, and risk)
Automating end-to-end processes and aiming for straight-through processing
Automating individual tasks on the desktop with attended automation
Imagine if we could make every employee 5% more efficient by just automating their local individual tasks? This extra capacity would result in employees handling more claims, more effectively.
The result: better customer experiences, more capacity for a growing business, improved retention, and new insights. These insights can lead to material cost savings, e.g., new preventive measures for loss mitigation.
I believe that claims operations will unlock digital transformation for the insurance industry. And insurance leaders can reduce claims operational costs without sacrificing the customer experience. Think about all the intersection points on the claims journey that can change from bad to good, or good to superior customer experiences.
If insurance leaders focus on straight-through claims processing, they are problem solving at the root of systemic challenges. They are also simplifying their legacy footprint with a focus on customers. Long, fragmented processes, silos, and regulation slow down claims processing and drive up the cost of claims.
A seamless claims process lowers operational costs and reduces claims processing times. It also dramatically improves customer experiences.
New to RPA? Learn the basics in our article What is Robotic Process Automation?
The efficiency game changer comes from automation getting the right information to the right person at the right time.
This translates to quick decisions and quick next best actions. The many legacy stops and starts are eliminated, so there’s loads of time freed up. Instead of employees gathering data on the customer and policy details, they are freed-up to focus on servicing the client.
Most insurance providers currently use several systems to get a single customer view. Because RPA sits across the top of all of these systems, integration has suddenly been eclipsed by automation.
Automation pushes and pulls data from one system to another as required by a process. That results in the right person getting information quickly and accurately. Whether that person is an underwriter, claims handler, fraud investigator, compliance, or auditor. That’s what RPA’s power is about, i.e., a connected and single customer view from ‘the front door all the way to the back office.’
RPA provides a ‘glue’ function because it supports the development of this ecosystem. And new and evolving capabilities are a material part of the claims transformation picture. There’s different technology that’s been around for a while, including chatbots, and optical character recognition (OCR). You have probably tried these types of technologies on specific opportunities or problems. All the tools are needed in combination to solve insurance problems.
Software robots orchestrate the process of ‘calling’ on these tools and functions as needed on task and process completion. This allows an enterprise to reuse the RPA capability and components at scale.
Learn more about common RPA insurance use cases.
And RPA capabilities are evolving. In fact, automation technology is at an inflection point. Automation is progressing from rules-based robots to robots that handle increasingly complex tasks. These more complex tasks aren’t necessarily codified by rules. This non-rules based orchestration is what some are calling intelligent automation.
What does it mean when we bring next-generation robots into the heart of claims processing?
Let’s assume you have a vision to automate the end-to-end claims process. First, you’ll need humans in the loop to make important decisions, check robot decisions, and oversee the process. From there, here are three common scenarios that RPA frontrunners are pursuing:
1. Integrated single-customer view
Objective: Connect multiple sources of data from underlying systems quickly and efficiently. Push and pull data as required to create an integrated, singular customer view.
Use case: A user interface for a contact center agent or a claims handler connected to multiple systems, e.g.: claim, policy, customer relationship management (CRM), know your customer (KYC)/sanction, complaints, policy wording, history of communication, etc.
Automation capabilities needed: Ability to build a dynamic user interface and integrate with an artificial intelligence (AI)-based capability that can recognize screen elements in any application, irrespective of technology or environment, even virtualized ones with high accuracy.
How it works: This AI-based capability uses a mix of machine learning (ML), OCR, and other technologies to enable RPA robots to automatically recognize on-screen elements within a user interface.
Skills needed: RPA team, contact center subject matter expert (SME), and IT knowledge of your IT estate.
Imagine how you can extend this capability for use with underwriters, customer support, complaints officers, and collections.
2. Unlocking unstructured data and enabling ML and AI
Objective: Optimize the access and utilization of unstructured data to reduce re-keying, gain more insight, and make better decisions.
Use case: Claims triage, e.g., email to get the right information to the right person for the next best action.
Automation capabilities needed: Text classification and entity extraction using ML models and OCR.
How it works: Robot capability can be augmented with additional skills using ML models that handle the challenge of high volumes of unstructured data and documents like emails or forms. These have been scanned and the text extracted using OCR, etc., that are not consistent and have a morphing structure depending on the use case.
Skills needed: Claims SMEs, data scientists, and data to teach the models how to distinguish claims types, policy numbers, claims numbers, etc.
Imagine how you could triage other correspondence such as complaints, renewals, and amendments much earlier and without errors. You’ll realize dramatically reduced waiting times and execution speeds.
3. Predictive analysis on claims for fraud prevention
Objective: To pre-scan a notice of loss claim to ensure it is a valid claim, applying risk factors prior to a claims handler being assigned or a settlement being made.
Use case: A fraud robot is applied to a motor vehicle or property claim to:
1. Collate data on the loss, claimant, policy coverage, history, etc.
2. Validate the claimant and the policy
3. Score against known fraud risk factors
4. Present to the claim file, now created and saved, for handler assessment or proceed to settlement
Automation capability needed: Use document AI and OCR to extract relevant information.
How it works: The software robot scrapes information from emails and forms, collates data from integrated policy and/or claims systems, and third-party data using APIs or AI-based computer vision to determine the validity of the claims, train your own model to apply risk factors and learn, and use custom models to manage and deploy your model.
Skills needed: RPA team, fraud SME, and IT knowledge of your IT estate.
Imagine how this use case becomes part of your whole triage process. Your low-value, high-volume claims could be handled by a robot and your skilled resources could spend time on handling the exceptions and oversight. This, integrated with a feedback loop to a fraud ML model, can learn or improve the risk factors. You will start to see that using this capability to solve insurance problems right at the start of the claims process can be reused across your company.
I am very lucky to be building these ecosystems with some of my customers and our great team at the UiPath Immersion Lab!
Get a behind-the-scenes look at a UiPath Immersion Lab.
Insurance CEOs are skating to where the puck is going. In reality, they haven’t had the technology to automate end-to-end and truly up-level their claims business until now.
For those readers who haven’t started yet, it may be intimidating, but the reality is everyone is just getting started.
To summarize the strategic approach discussed in this blog, my best advice is:
1. Start with claims processing
2. Use RPA to connect and integrate your legacy estate to provide that single customer view. This can apply to processes which support underwriters and call centers.
3. Be purposeful about breaking down silos and building end-to-end process automation
4. Develop your team to capitalize on your data. This includes the right skillsets to unlock the power of your data.
Before you know it you’ll be into your RPA + AI journey providing better customer experience and freed up capacity to support your growing business.
Want to learn more? Get your complimentary copy of our latest insurance white paper Automation for end-to-end claims processing: How design thinking combined with automation technology pushes insurers to evolve.
