There are many opportunities for startups to help propel the traditional banking and insurance worlds into the modern age. However while in my previous post I mentioned I’ve seen a lot of impressive fintech solutions from startups, at the same time there are still some glaring unsolved problems industry-wide.
This might be due to the problem being genuinely hard to solve, or industry outsiders being unaware that the problem even exists to solve.
Whatever the case, if these startups existed in Asia, I see an opportunity and would gladly participate in a seed round!
The Open Banking movement in the UK has resulted in an explosion of exciting new fintech startups. Having a standardized set of secure banking APIs means that startups can build engaging customer propositions, customers have greater transparency and choice of providers, and banks can innovate faster.
Subsequently, most financial institutions have caught on to the idea that in the future they must exist as an API. If all their products and services were API-based, not only could their own developers build on top of it faster but integration with partners can happen faster and more predictably. The benefits are self-evident.
However, for big traditional organizations, this move from legacy systems and often hard-coded integrations to a suite of modern APIs across their entire service offering represents a tectonic shift in terms of digital transformation.
A complicated matrix of legacy systems is the norm at traditional enterprises
It will take some institutions years to do this and many will fail along the way, spending billions of dollars in the process.
Additionally in Singapore, the MAS stance is to take an an organic approach towards open banking, meaning many Singapore institutions are being left to their own devices when it comes to API strategy.
There's an opportunity here for startups to come in and build modern middleware that sits on top of the legacy systems to dramatically reduce the time to launch new APIs to weeks, where previously it was years.
This startup would be part product, part service-based and specialise in integrating with financial systems and working with large financial enterprises. If no startup emerges to solve this, I would expect a company like Palantir to move into this space sooner or later.
There have been two interesting thought experiments recently on the subject of insurance claims. One is this video from a group of consulting companies showing a proposition that is pretty much achievable with today's technology.
This shows a vision of the kind of seamless claims experience that customers want - in their moment of need, the insurance company just takes care of everything with almost zero involvement or effort required from the customer.
The other thought experiment goes one step further but explores an area just as important as seamless customer experience - claims verification.
Black Mirror S04E03 - Crocodile
In the episode titled Crocodile of the series Black Mirror, an insurance claims inspector uses a device to scan a person's memories to corroborate their claims or uncover facts from witnesses.
Insurance is a business based on trust so most claims are paid without any sort of drawn out investigation. However, at the same time an insurance company needs to make sure they don't pay false or fradulent claims - payments of fradulent claims negatively affects honest customers, as premiums will need to rise in response to increased claims activity.
So the technology in the show solves a real problem. The present day process of validating claims / loss adjustment is largely a manual process. Pieces of paper fly backwards and forwards between customer, agent and insurance company in what is essentially an exchange of data. Police reports. Purchase receipts. Medical records etc.
While the technology to scan somone's memories is off in fantasy land, the technology to request, aggregate, overlay and analyse multiple data points is readily available in the present day.
AXA for example, already offer instant travel delay claims because it's relatively trivial to overlay customer-provided flight purchase data with publicly-available flight delay data and disburse funds accordingly.
What if you could productise this kind of instant claims validation across multiple types of products? Motor, Home, Health etc. Helping insurance companies pay claims faster and more accurately strengthens their relationship with customers while at the same time reduces fradulent claims payments and operational overhead - it's an area where a startup could add significant value.
If you’ve been to the bank recently, you probably had to fill in a paper form of some kind. A tremendous amount of customer operations at financial institutions are still paper based.
Banks and insurance companies know this is inefficient but it's incredibly hard to reassemble the engine while the car is moving.
Account opening forms, account renewal forms, change of address forms, change of beneficiary forms and so on. The sheer volume of different pieces of paper like these means that as above, to go completely digital is a tectonic transformational undertaking and in some cases would years to achieve.
The key to solving this is computer vision and machine learning.
Digital form tech is a well-established segment with companies like Typeform. There should be a startup that combines digital form tech with computer vision to help financial institutions quickly digitise their estate of paper forms.
Imagine if you could take a photo of a paper form and a computer vision algorithm assembles a digital version of the form instantly. A bank or insurance company could go paperless in a matter of weeks or months, not years.
This startup could then expand into other industries famous for their endless paper forms such as government, healthcare etc. Executed well, I would expect such a startup to eventually be acquired by Google as it aligns very strongly to Google’s strategic mission of “organizing the world’s information”.
Relationship managers are the people responsible for servicing the bank's high value customers. The majority of customers of the bank won't have an RM. This is not to say the bank doesn't care about these customers, it's just that the bank cannot service this larger number of customers in the same way, without it being very expensive.
Enter Artificial Intelligence.
Ex Machina, 2014
Eventually an AI-based RM will be able to do all the things a human RM can do, such as:
But it will be able to do this at a mass scale, in a way that is tailored to each customer, at a fraction of the cost of hiring the same amount of human RMs.
There will probably always be a place for the human-to-human relationship in the high net worth segment, but if banks want to reach wider groups of customers then AI is the answer.
Note that I'm not talking about some trivial, conversation-branching chatbot. We’ve already been through the chatbot phase and they have all failed miserably. Customers hate them.
I'm talking about a Turing test-passing AI that has access to your financial data and deep knowledge of the financial regulatory environment. It would be a brilliant marketing move for this startup to make the AI sit for regulatory exams and pass each one of them in a matter of seconds.
Startups who are external to the financial industry and are looking for interesting problems to solve in this space can refer to this set of MAS Problem Statements from 2017. These were collected from various institutions are compiled by the MAS for the Fintech Festival last year and are a great place to start looking for problems to solve in the financial industry.