Most of us have seen sci-fi film Terminator, a story of man and machines travelling back in time to alter the past and affect the future positively (depending upon your perspective).
The film’s back drop is Skynet, an “artificial intelligence” defense network that will become self-aware in the near future and initiate a nuclear holocaust. The time traveler’s journey back to 1984 from 2029; a date not so far off in our future.
So, what exactly is AI, and on the insurance battlefront, can it help us fight insurance fraud?
AI, sometimes called machine intelligence, is where machines demonstrate a form of intelligence and a level of awareness. It has been defined as “a system’s ability to correctly interpret external data, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation”.
In terms of identifying fraud or potential fraud our industry has long relied upon Risk Indicators, Red Flags or Fraud Identifiers. Such systems, automated or otherwise, are simply a set of experienced-based rules based on warning signs. A typical example would be a claim submitted only one day after Policy Inception.
We have learned much in the past 20 years and now treat “Risk Indicators” with a more open mind. Our industry now has many more tools which in turn support our fraud-related research and the risk in comparison to the identified Risk Indicators. Claims that subsequently fail our Risk Indicators and research are subjected to further scrutiny.
Despite concerted efforts, however, the battle against fraud is far from won. In 2016, insurers detected 125,000 dishonest claims valued at £1.3 billion. It is estimated a similar amount goes undetected each year. This is why insurers invest at least £200 million each year to identify fraud. Current estimates are that our industry has a 70% strike rate on proving fraud, which is to say that 30% of genuine customers are adversely affected by our efforts. Clearly, we need to find a better way.
Surely AI adapted to take over the decision-making process of fraud identification will improve our ability to prevent and detect fraud – assuming our data capture is sound? One would presume it could identify suspicious claims factors and cross match the data subject in multiple areas at a speed and level of accuracy far beyond human capabilities.
However, there is one element of fraud detection that AI cannot import – cynicism (a significant part of “natural intelligence”). We’ve all felt it, culminating in a furrowed brow. Something is not right. Perhaps we need to physically visit, to see, smell, listen, touch and acquire a taste for what we are being told.
AI is not Skynet, it will not (for the foreseeable future at least) be “self-aware”. Further, the insurance industry’s ability to capture data requires improvement.
To quote Donald Rumsfeld:
“There are known knowns; there are also known unknowns; but there are also unknown unknowns, and if we are to look through history it is the latter category that tend to be the difficult ones”.
It is a known that fraudsters patrol and exploit the gaps. Another is that fraud is a displaced crime; if we are weaker than another target within the financial services industry; we will be attacked by that displaced fraudster. Our known unknown is the fraud within our industry we are missing. And then there is the frightening prospect of the unknown unknowns.
So, how do we capture all fraud without affecting service to the genuine customer?
There is no silver bullet for missed fraudulent activities. At a high level, we must improve data capture and analytical abilities as well as communicating comprehensively across our own and other industries. AI will, without doubt, play a part in this.
At Crawford, we have been on a mission to address each stage of the fraud process: prevention, detection, handling, and sanctions. We need multiple answers to fill current gaps. Data is key, as too is the ability to automate, compare, analyze and finally handle our customers with the appropriate speedy response, with complete empathy and understanding.
In addition to the usual cross product services, we are developing our data analysis with a range of cutting-edge data led innovation tools and our intelligence service, Insight. We can address every level of claim in a commercially sensitive way, having developed WeGoLook for a pure eye on fraud. WeGoLook is, without doubt, the most innovative mix of technology and the human touch, delivering immediate data (24/7 365) to our clients so as to provide best-informed decisions. And Crawford Legal Services is now attacking our fraud caseloads with a complete arsenal of sanctions offerings, including the pursuit of proven fraud with Tort of Deceit cases.
In the immortal words of Arnold Schwarzenegger; “I’ll be back”.
If there are any questions or enquiries about our full list of Fraud Service offerings, please contact a Crawford expert.