Do you see yourself nodding along when people mention "blockchain" or "machine learning"? Are you still unsure what's on (IBM) Watson? I am here to lend a helping hand.
"Insurtech" is a nebulous concept broadly referring to innovative technology in the insurance industry. Comments on insurtech are often littered with references to specific computing concepts which individually are pretty straight-forward. However, when used in conjunction with one another with no further explanation, insurtech jargon can sound much more complex than it really is. I explain some of the key terms below.
What is it?
A "blockchain" is a way of distributing information between a network of people in a way which is traceable, transparent, and cannot be retrospectively amended.
In a blockchain each "block" (chunk of data) has a unique signature and contains information. This information can only be amended by creating a new block which links to the previous, creating a "chain". Other rules can also be added to the system, such as incorporating the names of parties who amend the information into the blocks.
Information in a blockchain is saved on a "distributed ledger", a copy of which is given to all participants in a (private or public) network. Consequently, the blockchain is visible to all participants and, as blocks cannot be amended but only added, data cannot be retrospectively altered.
Blockchain can be used to reduce paper-based information, distribute information quickly to others, and reduce the risk of lost documentation or duplication of claims.
Information in a blockchain can also be audited or analysed to, for example, quickly find the number of claims under a policy or total amount paid under a policy. The auditability of blockchain has led to the FCA promoting its adoption.
What is it?
A "smart contract" is a contract based on a blockchain containing self-executing clauses. The blockchain element ensures the contract is tamper-proof and the self-executing clauses automate processes.
Smart contracts can provide visibility and certainty over terms in complex negotiated multi-party contracts, automate payments when certain conditions are satisfied, and provide an auditable trail of an insurance claim. It was recently used in a flight-delay insurance product and piloted in a multinational commercial insurance policy.
AI/Machine Learning/Predictive Coding
What is it?
Whilst there are technical differences, AI, machine learning, and predictive coding are often used interchangeably to describe a process of automating a repetitive task in a manner which allows 'experience' from each repetition to improve the overall process.
The concepts refer to a computer's ability to review a data set, understand how it should be processed by following instructions or comparing a human-reviewed data set, and then make predictions on that data, often in an iterative process.
Machine learning could be used to quickly recognise fraudulent claims or claims which can be fast-tracked, automate elements of a standardised claims process, or create chat bots to deal with queries (possibly as part of the claims or distribution process). Machine learning can also be used to help underwriters quickly process data on a risk and therefore price risks more effectively and competitively.
Most famously, IBM Watson uses vast stores of data to answer questions based on probabilities.
Data analytics/"big data"
What is it?
Data analytics and "big data" are also often used interchangeably. They have different meanings but both relate to describing new insights into data.
"Big data" refers to large datasets which can now be captured, far beyond previous generations of paper-based information and using much smaller memory devices. "Data analytics" refers to the ways we can analyse this data, which have grown exponentially over the past few decades.
New ways we gather data and its transmission to insurers has led to innovative insurance products and techniques, such as motor insurance by the journey, analysing historic solar data to insure solar energy output, and using facial recognition to estimate life expectancy.
So what does it all mean?
As insurtech continues its growth throughout London, and the rest of the world, these concepts will help to make insurance cheaper, more efficient, and more innovative for everyone.