Osman Khurshid, Head of Solutions Consultancy, Synectics Solutions
For the financial services and insurance markets, the shift to a more automated and digital customer experience creates savings and efficiencies - business can be written quicker and customers receive a fast, smooth experience.
However, these digital advancements also cause the same industries a new headache. Sales and Marketing teams are pushing for a shorter apply processes to avoid drop off, (and therefore often ask for less up-front information). They want an ultra-slick sales funnel. Compliance, Fraud and Risk Teams on the other hand need to ensure that the necessary credit, risk and identification checks are in place to guarantee that they comply with Know Your Customer (KYC) regulations.
Unstructured data is information that either lacks a pre-defined data model or is not organised in a pre-defined manner. It is typically text-heavy but may contain data such as dates, numbers and facts. Unstructured data typically makes up 80% of an organisation’s data, and this is growing at a rate of 55%-65% each year*. Structured data such as names, addresses, and phone numbers are usually easy to capture, store and maintain digitally, whereas unstructured data has severe limitations.
Extracting data from hard copy documents such as passports, utility bills, invoices and claims forms has historically been a purely manual process, carried out previously as part of the customer on-boarding process, or during an investigation or claims process for insurers. It robs staff of precious time and errors naturally result from manual, human input.
Thankfully, new tools are emerging that enable financial service providers and insurers to extract ‘locked up’ unstructured data, digitally, allowing them to leverage intelligence and vital information held in hard copy documents. This helps organisations to enrich and enhance the information they hold about a customer, financial application or insurance claim.
Adrian Webb, Marketing Director, 360Globalnet
Ultimately, the developments are enabling organisations to streamline their internal processes and improve their ability to access much richer levels of information when doing their fraud analysis - by having a greater depth of information at their fingertips – and not having to go through the laborious and time consuming process of sifting through this intelligence manually.
Mark Hobart, Director, Infoboss
I’ve looked at two cases, where extracting unstructured data into risk mitigation solutions to form a broader picture of a customer has enabled businesses to lower risk, increase efficiencies, reduce losses, reduce indemnity spends and improve the experience for their good customers.
Claims reserving
The UK insurance market is the largest in Europe, the fourth largest in the world**, and it continues to grow and adapt with the emergence of Insurtechs and the changing behaviours of consumers. Competition is high and insurers are finding it increasingly difficult to remain profitable. By Q3 2019 the average private motor premium had fallen from an all-time high of £491 in Q4 2017 to £468. In addition, 98.4% of motor claims were accepted***.
One area that feels the resulting heat and pressure is the Claims Reserve Teams. Accurately reserving the value of a claim is becoming more and more important for insurers. By understanding the true cost of a claim, companies can ensure that reserves are accurate and premiums for customers are competitive.
Historically, Claims Reserve Teams have relied on the experience of Claims Handlers to understand and identify the key warning signs of whether a claim poses a risk of costing significantly more than was originally reserved. The problem is that this process is predominantly a manual process; it also relies on the knowledge and experience of individuals being passed to new members of staff, their knowledge and processes being consistency applied to estimate the reserves.
Insurers are now utilising machine learning and predictive analytics tools to help them foresee the types of claims that are likely to exceed the allocated reserve. In addition, to enhance the accuracy of their predictive analysis, insurers are increasingly using unstructured data mining services to identify key phrases, words, behaviours, trends and patterns associated with those claims that result higher pay-outs for claims than was originally forecasted.
Fraud prevention
The underpinning of any effective fraud prevention solution is relevant, timely and accurate data that is easily accessible by the business. In the past, fraud prevention systems have taken advantage of structured datasets such as National SIRA and Cifas as well as third-party intelligence sources. However, there’s often far more intelligence that fraud investigators can utilise from their own unstructured data. This is usually held in invoices, statements, utility bills, payslips, claims forms and identification documentation. By harvesting this information and integrating it into fraud prevention systems, companies can employ complex data analysis techniques which enables them to gain a much more informed, rich picture of the levels of risk associated with applications or claims.
It’s no secret that collusion is rife in the insurance industry – in particular amongst third-party enablers such as solicitors, brokers or service providers. One example of this is vehicle repair fraud, where repair garages are involved in fraudulent claims by falsifying or overinflating invoices for work carried out. In some cases, invoices by-pass the insurance companies and are sent directly to solicitors to avoid any manual checks being made.
Prior to the introduction of unstructured data mining capabilities, those involved in the collusion would know that information locked-up in invoices, receipts and other paperwork sent via a solicitor to an insurers would not be manually entered into case management or fraud systems and would therefore often go undetected.
Now, by unlocking the information from the unstructured data document (invoices, receipts etc.) the insurance companies would be able to more quickly identify the false information and induced incidents that were linked to a fraudulent claim. Unstructured data tools can ingest data from incoming documents within seconds and make the intelligence available within fraud prevention systems to integrate within fraud investigations – removing the need for the manual entry of the information.
How synectics solutions can help
Synectics Solutions has recently partnered with both 360Globalnet and Infoboss to provide its clients with unrivalled access to unstructured data mining technologies, which can be integrated within SIRA. 360Globalnet works with insurers to offer them digital technology solutions that will help revolutionise the claims process. Its product, 360Retrieve allows previously ‘locked up’ information, held in documents to be made fully searchable and therefore integrated with fraud and claims management systems. Infoboss works with organisations to help them better automate the management and processing of data. Its data mining capabilities allow the collection, storage and indexing (in a searchable format) of electronically held, unstructured data.
For more information about unstructured data and how these services can help enhance your risk mitigation and claims management process please call: 0333 234 3411
*www.datamation.com/big-data/structured-vs-unstructured-data.html
**Swiss Re, Sigma No3/2019
***www.abi.org.uk/globalassets/files/publications/public/key-facts/abi-key-facts-2019.pdf