Adoption of impactful and mid-term digital transformation is revitalising insurers‘ biggest asset – data. Machine learning has long passed the stage of feasibility experimentation and has ’snuck into‘ the whole insurance value chain. Data science on the whole continues to transform business processes into machine learning systems.“
So can we hope that insurance is no longer disinclined to change? Or has Covid-19 shifted insurance transformation efforts up a gear? Has the future of claims arrived prematurely, as some commentators are saying? Over the last few years, it’s been said that we’re at a crossroads for insurance. A feeling of deja vu persists. The industry is yet again at a crossroads, but this one forced upon it by a global emergency. Thomas provided some examples of where cognitive claims can be applied. Particularly, for insurers looking to fortify their business against the systemic vulnerabilities the pandemic revealed.
Let’s explore:
Customer interactions become actionable data
1: Communication channels are fully digitised and omni-channel service (be it desktop, mobile, phone, or in person) is becoming the norm.
2: Customers engage in the channel that is most convenient to them. All associated data should be available to the carrier.
3: Immediate understanding of the coverage allows the system to guide claimants through the process. The claim submission is personalized to the loss type, line of business and claimant’s history.
ML can now support cognitive claims systems
1: These machine learning systems support the adjuster in identifying key risks to a resolution from large volumes of claims handling data.
2: They enable the expansion of fraud screening capabilities (omni:us is partnered with FRISS to this end).
3: Selecting the ideal claim adjuster for a task, according to claim complexity and adjuster track record is a facility with growing traction in commercial lines.
4: Governance throughout the claim phases is simplified as each phase of claim handling can be converted to insightful data. This info is accessible to claim operation teams in real-time.
Loss data is augmented with contextual info
1: When FNOL submissions are proficiently understood, the attention shifts towards the automated understanding of how the claim happened.
2: Adjusters receive an assessment of all identifiable circumstances. Manual research workload is reduced.
3: The digitized claim is cross-referenced against the claimants history. Photo & video material is analyzed and validated with sensory data where available.
Gap identification and clarification
1: Any uncertainties to the claim’s validity/completeness are ironed out by identifying gaps with the cognitive system.
2: AI-enabled coverage understanding combines all available information and provides in-depth contextual analysis with close reference to the terms & conditions of the insurance product.
3: The adjudication of a claim, together with the identification of applicable rejection reasons, is applied to the claim handling data.
4: Risk assessment scores are used to determine who should review the claim and recommendations are given on how to operate effectively while reducing claim leakage.
Above are just some of the examples of applied AI in claims. It’s evident that real difference can be made to both customer and employee experience with some adjustments to existing systems. While implementing such changes are not by any means straightforward, they can be made smoother with the help of experts who understand both insurance and artificial intelligence.
____
Art: wenmei Zhou
