In an ideal world, technology should be an enabler. It should empower an underwriter to do more, see more and understand more about the risks that come in through the door. Technology should enable the underwriter to be free to focus on applying judgement to complex risks without having to spend hours and hours sorting the wheat from the chaff.
This is where algorithmic underwriting comes in.
Although a relatively new term, it's meaning is simple: the use of artificial intelligence and machine learning to help sort, assess, and match risk profiles with an underwriter’s appetite and interest. Removing, in other words, the need for an underwriter to spend hours poring over a risk submission, validating any data, and manually comparing this to existing policies and wordings before deciding if it is one for that insurer, or not.
Only policies that meet the carrier’s specific criteria will be presented to the experienced and knowledgeable underwriter, avoiding the need for them to spend valuable hours on manual policy comparisons and clause analysis.
Algorithmic underwriting is already advanced in some areas of insurance, particularly in simple, commodity risks such as personal motor insurance. As technology develops, and digitalisation of data at both the back and front-end advances, it is, and will be, increasingly applicable in the more complex commercial and specialty lines.
For algorithmic underwriting to work in practice there are a number of key enablers that need to be in place.
Incoming data about the client and the risk needs to be distilled, harmonised and organised. Fortunately, the onus no longer rests on the broker to the heavy lifting of this complex data. Modern AI-based platforms today can quickly apply structure, and enable details to be automatically validated by third-party providers.
The insurer’s risk appetite, criteria and tolerances need to be pre-set, to enable this to be applied to and married with the provided information.
Policy documents and associated clauses need to be digitalised to enable automatic mapping against incoming data — and not just on PDF or scanned paper documents.
If this is all in place, then algorithmic underwriting will enable an insurer to use automated decision-making to assess, filter and in some cases agree quotes on what are relatively complex commercial risks. It means that unsuitable risks are identified quickly, and automatically, helping save both the underwriter, and the broker, valuable time.
Only policies that meet the carrier’s specific criteria will be presented to the experienced and knowledgeable underwriter, avoiding the need for them to spend valuable hours on manual policy comparisons and clause analysis.
Clearly, the more information that underwriters receive, the more data can be analysed by the system, the better the decisions.
The main benefits of algorithmic underwriting are:
Better risk selection: Automated, intelligent filtering of presented risks against pre-set criteria;
Speed and service: Faster underwriting cycle, and the efficient rejection of unwanted risks;
Expertise focus: Free up underwriters to spend less time on basic administration and more time applying their experience and insights to complex risks;
Future-ready: Digitalisation of processes and the use of machine learning enable rapid update of new technology;
Broker service: Fast turnaround times, saving brokers time and providing them with clarity.
Many commercial lines have been slow to benefit from the adoption of new technology, with the view that only skilled, experienced (human) underwriters can make accurate decisions on complex risks. This is true to a degree, but with algorithmic underwriting supplementing the skill of underwriters, commercial insurance businesses can quickly weed out the ‘wrong’ risks, allowing a laser-like focus on those that require the expertise and knowledge of hands-on underwriting staff.
The benefits speak for themselves: algorithmic underwriting has the power to cut through the complexity of commercial lines and allow for more efficient, fast and accurate risk selection. Not only can underwriting teams save time, but they can also quickly gain a more holistic risk profile of a customer, and ultimately make more profitable portfolio decisions to drive growth.
