The potential of e-trade and the COVID-19 pandemic
In essence, the Insurance market is a people-led business. Through thorough networking, leadership and avid discussion on continuous improvement this sector has maintained its global position and resilience. However, times change and I can personally see that first-hand, during my first 20 years in the London Market, whether that’s queuing in Lloyd’s for hours on end or receiving ‘quotation terms’ on the back of an envelope or even on a cigarette packet, when the underwriter was tracked down ‘networking’ after hours.
The novel COVID-19 pandemic has stimulated a greater reliance on digital delivery and many insurers have found that their own e-trade platforms are not robust enough. This issue may have only been extrapolated during lockdown. It is anticipated that the use of e-trade is set to increase with many insurers exploring placement of further complex risks electronically. However, a fundamental weakness in any digital delivery, irrespective of platform is the introduction of referrals. Referrals cause a hard stop in an otherwise seamless digital journey and can only be overcome once there has been some form of manual intervention.
Creativity benefits in challenging times
We, at Softelligence, are incredibly fortunate to work in an environment where creative freedom is actively encouraged. Our CEO, Adrian Bildarus, once said that “the insurance market proposition has not changed, but how we deliver it has”. These words have resonated with me since, inspiring us to solve real time business pain points with technological solutions – but only where appropriate. As an ex London Market professional, my intention is to turn this time of uncertainly, into one of momentous opportunity.
Following a recent study of a London Market provider, we identified that 42% of this insurer’s policy transactions triggered a referral. This introduced an unintended ‘hard stop’ within their digital broker journey. The disruption required a manual intervention in order to continue, which significantly impacted operational efficiency but also the integrity of the insurer’s broker self-serve proposition. Acting on the task of finding a solution, we have been able to utilise our unique Machine Learning insight and engineer a resolution that would save up to 938,804 business minutes over a period of 12 months.
What is Machine Learning?
Machine Learning is a form of artificial intelligence (AI) that enables a system to learn from data rather than through explicit programming. Machine Learning (ML) uses a variety of algorithms that iteratively learn from data to improve, describe, and predict outcomes.
How can the Softelligence Machine Learning solution address this business problem?
We believe this solution can bring value to Brokers along these lines:
|Reduce insurer dependency||The reduction in referral handling will lessen the need for Insurer engagement on e-trading platforms and provision time to be redeployed into growth activities and a fulfilled seamless digital journey.|
|Consistent underwriting||By applying and aggregating referral acceptance according to each unique risk profile, underwriters will be able to automate referrals with the help of a solution that employs Machine Learning.|
|Remove manual disruption||The current hard stop created by referrals will gradually dissipate with the application of Machine Learning insight.|
|Simplicity for clients||Enabling a digital journey without disruption will allow brokers to trade in real time.|
|Reduce operational costs||The current referral load has a substantial implication on processing times. The removal will provide the ability to redeploy time and costs into other value add activities.|
|Minimise the likelihood of errors and omissions||Our Machine Learning capability will reduce the exposure of human error lessening potential omissions|
|Enhance transparency||With evolving regulatory transparency, brokers will be able to evidence Treating Customers Fairly (TCF) conduct at all times.|
At the same time, the solution brings important benefits to the Insurer as well:
|Save time||The removal of high volume, low risk referrals will provision substantial time savings that would allow time for underwriting activities to be up scaled into the review of complex perils and risk management activities.|
|Risk profile clarity||The enabling of the Softelligence Machine Learning solution will further enhance the Insurer’s ability to begin risk segmentation in line with underwriting rate appetite.|
|Reduce operational cost||Scaling down the dependency on manual intervention will substantially reduce the products’ operational costs.|
|Fairness for customers||With evolving regulatory transparency, insurers will be able to evidence Treating Customers Fairly (TCF) conduct at all times.|
|Referrals of the future||Softelligence’s Machine Learning algorithm will only become stronger with enriched datasets and curated data imputation strategies, ultimately being able to underwrite 100% of referrals with optimal accuracy|
Challenges in the referral activity of a London Market provider
On inspection of the referral activity from a leading London Market provider, we were able to uncover a number of anomalies despite similar risk profiles and an excessive number of referrals.
At a glance we could confirm that only 1 in 2 quotations completed the digital journey uninterrupted. Even more so, 40% of the day is taken by referrals.
Our statistics initially suggested the insurer in question was attentive and performing well, which was echoed through a fair average of approved referrals within 32.21 minutes. On an individual basis this does not appear to cause a significant delay in completing a customer journey, but when aggregated across 6,740 quotations, the underwriting time loss is intensified. Exploring this matter further, we found clear commonalities in accepted referrals including risk profiles, referral reasons, and premium bands. This presents an opportunity to save underwriting and client time through the e-trade platform but also the ability to augment Machine Learning outcomes on risk behaviour and claims performance. Furthermore, given this volume of disruption in time and cost, and considering a national average insurance broker salary is estimated at GBP 29,213, it could be considered that the operational cost of the referral disruption on salaries alone is GBP 215,924 – which is a light estimate.
While these figures are alarming, we fully acknowledge and accept that we have defined these commonalities and they will change from insurer to insurer but the application remains consistent. Therefore, we actively encourage the full engagement of a collaborative workshop to further investigate these risk profiles with the support of sector knowledge.
Softelligence’s Machine Learning Solution
It was estimated during a recent TINtech event that 80% of underwriting roles (non-Lloyd’s) will be replaced by Machine Learning in the next 20 years. While the pace of innovation is varied it does support the need for Machine Learning as its real time capability is fast becoming superior.
How will our Machine Learning solution work for you?
|Discover||Collaboratively analyse the data together and discover whether further imputation strategies can be adopted.|
|Automated referral||Enable the ability to automate referral decisions based on legacy underwriting behaviours|
|Data lake||Introduce a data lake to enable unstructured data to be held and uncompromised in hierarchy, elevating any possible restrictions.|
|Continuous training||Like an employee, Machine Learning can be continuously trained to assign new referrals to an existing segment, further improving its capability with every transaction.|
|Enhance further||Using such recommendation results, implement further components that can automate complex pricing and/or assist underwriters by driving quicker referrals.|
|Real time results||Like an employee, Machine Learning can be used to assign new referrals to existing risk segments once they have been created, reinforcing people to continue the process naturally.|
Next steps and getting started
Our implementation journey is unusually small for projects of this scale; however, this is a direct result of the considerable production Softelligence underwent in readiness to solve this issue. This ability provides insight into tackling those referrals that carry common trends and showcasing those that require further data to enable automation. Machine Learning is a journey.
It is our belief that insurers looking to optimise their e-trade platforms away from manual intervention should explore an engagement with us. The scale of this capability is able to accommodate multi product, multi insurer and a diverse trade demographic. Should you look to implement, this can cater for high volume SME business driving efficiency throughout your business.
Less dependency on human intervention reduces the transactional costs. These savings can be measured and assessed whether to pass on this reduction through rate or reinvestment into IT initiatives becoming stronger with every transaction.
If left unaltered, as shown from this case study the current challenges will only set to increase both in volume and risk complexities. Our Machine Learning capability is one of many technical insights which can support an alignment to risk, actuarial and claims performance ensuring a robust delivery.
Achieving your ambitions
At Softelligence, we have maintained a proven track record in minimizing disruption to the activities within the London market. Our digital transformation consultants are committed to serving your ambitions, and are prepared for the implementation of solutions across a diverse sector. We welcome the opportunity to discuss this solution with you, while paying close attention to pain points that you specifically wish to investigate. We hope you will agree that the potential of this opportunity is limitless and will drive seamless journeys into a new realm of possibility.
We are keen to get started. Are you ?