20 Dec, 2020 • 3 min read

InsTech London - How Re:infer is Helping Insurers Become InsurTechs

In 2020, Re:infer joined brands such as LV, Lloyds, Deloitte, and AXA as a member of Instech London.

An organisation renowned for identifying and promoting the best technology, data, and analytics innovations within insurance and risk-management. Since their establishment, Instech has shined a light on the most promising insurance innovation. 

Their co-founder Robin Merttens recently loaded up a video-call with our CCO Stephen Mackintosh to discuss Re:infer’s role in that innovation.

Below, we have summarised the key takeaways from the podcast, but we highly recommend you watch the full 20 minute session for more detail and a preview of the Re:infer platform in action.


How Communications Mining Works

Re:infer was launched out of the UCL AI Research Lab, home of illustrious machine-learning companies including Google DeepMind. Re:infer’s focus is on state-of-the-art natural language processing (NLP) and communications mining.

Insurance companies run on communications data, from email, chat and tickets, to annotations and CRM notes. The unstructured communications data is rich with process intentions, sentiment, and semantics as well as key information like policy numbers and trade IDs. 

Re:infer’s communications mining platform converts that unstructured data into useful machine-readable, structured data, enabling Insurers to automatically discover patterns in conversations — finding frequently occurring questions, requests, processes and themes. 

The zero-code interface empowers business users to teach Re:infer the specifics of your business, delivering value in hours, not months.


Re:infer’s Success with Financial Institutions

As Re:infer’s reputation grew, the world’s largest financial organizations took notice. When first examining the processes of a large investment bank, Re:infer analysed 15million+ emails across Post-Trade Operations, identifying bottlenecks, inefficiencies, risk events, complaints and exceptions, but also client themes and clear revenue opportunities.

Unstructured data is largely invisible to and unuseable by a bank’s real-estate. By extracting, measuring, and quantifying operational processes, Re:infer was able to identify those that the client should eliminate, augment, or automate. This allowed the bank to achieve levels of organizational efficiency that were previously considered unattainable.


Re:infer’s Success with Insurers

Following well-publicized success with Hiscox, insurance has become a major growth industry for Re:infer. 

Re:infer’s initial roll out with Hiscox was automated email triage. 97% of broker requests to Hiscox were incorrectly first routed to underwriters thanks to processes reliant on manual, individual action and admin. This manual workflow (or manual way of working) created significant intellectual waste across skilled underwriters. 

Re:infer’s communications mining platform allowed Hiscox to streamline operations, liberate senior underwriters from transactional repetitive tasks, and save the company hundreds of thousands of pounds in efficiency gains.

“Communications mining is elevating the broker experience. Re:infer is a capability, a tool that lives inside an insurer's operation, and really helps them to innovate.” Stephen Mackintosh, CCO, Re:infer

Re:infer’s platform has been proven to provide much faster resolution, efficient processing, and more gratifying results. Broker requests can be labelled, ranked, and filtered to discover which are receiving the worst experience, and by what measures that could be improved. Requests that were taking weeks, now take days.


Re:infer’s Client Profile

The clients with the most to gain from Re:infer’s services are large enterprises which send and receive high volumes of daily communications. Since the Lemonade IPO, NLP tools have begun to transform the insurtech sector. Thanks to SaaS solutions like Re:infer, large incumbents can embed those same technologies in their operations.

“We're not an insurtech, we're a machine learning company, but we help insurers become insurtechs, that's for sure.”
Stephen Mackintosh, CCO, Re:infer


Re:infer’s Platform

As individual messages are mined for vital details, the platform provides a real-time structured data summary that can be consumed by the broader IT real-estate.

The UI features helpful tools including reports, filters, and a tree map to proportionally represent communications. Together, these tools provide clients with a customisable “30,000 foot” overview with which to reveal the factors driving spikes in activity and optimise their operations.

“The whole UI is designed to put the business user in the driving seat”
Stephen Mackintosh, CCO, Re:infer

Re:infer’s zero-code platform is dynamic and intuitive. Employees need only point and click to add or remove labels, and the machine learning will automatically begin retraining in the background. 

Re:infer also has pre-built integrations for systems including Salesforce and 365 Exchange, as well as downstream connectors for workflow tools like Appian, ServiceNow, BluePrism, and UIPath. 

The goal is to make the link between unstructured data and action as seamless as possible without the need for technical training or bespoke build-outs. 

It is a SaaS application that can be played with, tested, and customised by the business for the business.



Resources Hub

The latest research, best practices and guides to delivering business value through Artificial Intelligence, Machine Learning, Data Science and Natural Language Processing.