Trust is the foundation stone of commercial insurance, but it is difficult to rely on it in the world of digital business. Howard Green explains the problem and suggests some solutions.
At the heart of the great London insurance market sits the historic London Institute of Insurance, its coat of arms inscribed with the words Uberimma Fides (utmost good faith). This is a reminder that our global business of commercial insurance has special origins and characteristics that we ignore at our peril as we move forward into the age of on-line transactions and digital business.
The ability to form binding contracts is one of the pillars of the commercial legal system. Because the idea originated to govern the trading of tangible goods, it was based on the doctrine of caveat emptor (let the buyer beware), meaning it was up to the buyer to inspect the goods and be satisfied as to their sound condition and quantity before handing over his bag of gold sovereigns.
But caveat emptor did not work for insurance, because you cannot inspect a promise, and the promise is in turn based upon representations that cannot always be checked, for example: "My ship is anchored off Virginia full of tobacco and cotton". Consequently, insurance adopted a new legal doctrine based upon utmost good faith, and it allowed for contracts to fail if the faith is broken through, for example, misrepresentation.
Uberimma Fides has importance beyond its legal implications; it also explains a lot about the way the insurance industry is shaped. Trust is vital to its workings, and it achieves trust largely by dealing with people it has become familiar with over time. In the paper-based world, trust and familiarity have become one and the same. But in a digital world this code of familiarity does not work. Firstly, digital networks expose us to myriad counter-parties outside our familiar circle, and they must be considered in order to remain competitive. Secondly, even people we know cannot easily be identified when we are transacting with them across a digital network. How do I know that the Bill Smith on the other end of the computer interaction is really who he says he is?
This issue of trust is vitally important. On the one hand, commercial insurance must speed up the adoption of computer methods if it is to remain competitive, reduce costs and provide clients with modern services. On the other, client confidence in the system will evaporate if there is a resulting erosion in the age old foundation of trust.
A parallel concern must be the clear balance between trust and regulation. Our foundation on trust has enabled commercial insurance (at least for large and complex risks) to stay relatively free from regulation. But these are changing times. Scandals such as Enron and WorldCom are swinging the pendulum towards greater regulation. The insurance industry's record keeping and trading practices are likely to be under deeper scrutiny than ever before. The inclusion of the London market under the regulatory umbrella of the UK Financial Services Authority (FSA) is an example of this.
Maintaining trust in a digital world is not a simple matter and involves a complex set of issues. These are more easily dealt with if they are understood and considered up front when planning an e-commerce strategy. It is vital that insurers and risk managers do not view trust as an IT issue. Trust is a critical business issue, and it must be part of the strategic plan considered by senior management.
One of the scariest things about the insurance industry today is its almost total reliance on e-mail as a means to interact, structure and form transactions. One broker recently commented that it has over 300,000 e-mails each day from its London office alone, and has no way to keep track or route them to applicable client or transaction folders. E-mail is a successful and useful communication tool but it is arguably the least trustworthy of any form of interaction – in fact it fails every trust test, as you will see. This risk is greatly increased by the equally addictive practice of exchanging vital documents by attaching them to e-mail messages.
Authentication
When making transactions across a computer network we cannot see or hear the counter-party. Thus we have no way of knowing for sure that they are who they say they are. How do we know if someone has stolen Bill Smith's password and is impersonating him? How can we tell if a hacker has stolen his identity? It is impossible (or at least not practical) to authenticate conventional e-mail messages. This creates two problems.
It is easy to imagine the opportunities that this opens up for mischief and fraud, particularly when conventional e-mail is used to structure and form valuable transactions. The worrying thing about insurance is that e-mail is increasingly the method used to negotiate and bind transactions of every kind.
The most reliable authentication is achieved in two steps. First, provide the counter-party access to messages or information on a device (a server) that you (or someone you trust) controls, rather than send information across the internet to an unknown destination (as with conventional e-mail). Second, authenticate the counter-party access to the device through one of three possible levels - a user ID and password (controlled by you), a digital certificate (a type of digital signature), or an authentication token (a key fob device with a dynamic password that changes every 60 seconds).
Collaborative commerce enables strong authentication because it works on the principle of granting access as opposed to sending. A risk manager, who might today e-mail a claims exhibit to a broker or underwriter will, using collaborative commerce, grant the broker or underwriter access to the message and document which is stored on a secure server. The broker or underwriter can access the server over the web but will need to authenticate himself either by user ID and password, or by the stronger methods explained above.
Using this model the risk manager can be certain that only the authorised broker or underwriter can access the information. Equally important, the risk manager will also have a record of when the access took place and what the broker or underwriter did with the document (changed or versioned it).
Certainty of electronic documents and messages
In commercial insurance we typically form contracts by exchanging large numbers of documents. Because these documents are often material to the contract, it is vital we are certain about, and can trust, the versions of the documents that have been exchanged. Today, the majority of documents used in contract formation are composed on PCs and evolve through multiple versions before and after they are exchanged with the counter-party. Yet PCs and the networks they are linked to were never designed to keep track of the intense document versioning typical of insurance.
Furthermore, documents are exchanged by attaching them to e-mails. In addition to the authentication and security risks, this has the effect of proliferating document versions. If I send an e-mail to Bill Smith with an attached document, we now have two versions: one on my PC and one on his. When these documents are being used to form contracts, who is to say which is the true version?
The harsh reality is that most brokers and underwriters have millions of document versions stored randomly in PC and laptop hard drives and network servers. To make matters worse, they keep a parallel set of paper records. It is increasingly difficult to reassemble records that are fragmented between filing cabinets, e-mail inboxes and outboxes and electronic documents stored on hard drives and servers. Look no further than the aftermath of the World Trade Center disaster to see the danger this poses – dozens of law suits, because companies found they had two copies of the same policy saying different things.
Of course, in spite of the dangers, the insurance industry is hardly going to throw away its massive investment in these systems, and users are not going to give up the benefits of e-mail and Microsoft Office. Collaborative commerce software has the ability to wrap around these deployed desktop systems, and restore certainty by automatically creating a detailed audit trail of every interaction, document version and transaction. Because collaborative commerce is built on the standards-based concept of web services, it can integrate with existing e-mail and desktop office systems automatically without actually visiting each desktop.
A key additional benefit is that collaborative commerce enables 'extended enterprise processes', meaning that risk manager, broker and underwriter can all be pulled into a single process where documents are formed collaboratively. This overcomes the problem of proliferating document versions.
Non-repudiation
A key element of trust is ensuring that the parties to an insurance contract have the same intent and interpretation. Many disputes now occur precisely because there have been conflicting intents. This problem increases as contracts become more complex and as companies broaden their activities to include financial risk hedging and alternative risk transactions. Clearly, no system will completely eradicate repudiation. Insurance contracts are complex and will always be open to interpretation. However, much can be done to ensure that the counter-parties are at least working with the same information and have a common record. Many disputes arise because there are conflicting records of what has been said about the risk, or what has been agreed in the contract. They are mostly preventable.
Using collaborative commerce to form contracts means that the counterparties are building a collaborative record of the transaction – a string of interactions and document versions exchanged between the parties. If the counterparties subsequently have access to this record and are prevented from ever changing it, this becomes a record of the transaction which is not open to repudiation.
Permissions
In the days when I was a broker you could get fired if you cut a $50 invoice outside the corporate accounting system, but nobody raised an eyebrow if you evidenced a billion dollar placement transaction on a beer mat. I am sure things have improved since, but this does make the point that 'permissions' (what employees can and cannot do) have never been a strong point in the insurance industry.
Collaborative commerce software provides sophisticated permissions structures, because it enables a unique interface to be rendered for each user. The administrator decides what tools and capabilities (permissions) each user will have.
Every company involved in commercial insurance has launched, or is building, a digital business venture of some kind. Evidence points to trust not generally being a key requirement in the design of these systems. This is worrying, because trust is fundamental to the risk business and vital to its future. It is important that business leaders appreciate the elements of trust in a digital world and ensure that they are a priority in any development.
Howard Green is president of Riskclick, Tel: 020 7436 5444, E-mail: hgreen@riskclick.com
SECURITY SURVEY
IT security training is woefully inadequate within 73% of organisations from the UK's largest employment sectors. According to a recent survey, almost three quarters of workers – from the accounting and legal sector through to the civil service – have never received any training from their employer on how to use the internet and e-mail at work to minimise network security problems. Further, although more then 80% of employees admit to being concerned by the risks posed to the network by viruses when they use the internet or e-mail, they are seemingly ill-equipped to identify and deal with potential threats, such as avoiding the temptation to open dubiously titled e-mail attachments.
The survey, commissioned by SurfControl and conducted by NOP World, analyses IT security awareness amongst 100 workers from each of the UK's six largest sectors (civil service, accountancy, legal, financial services, retail and manufacturing). Steve Purdham, CEO of SurfControl said that IT security training should be led jointly by the IT and HR departments and initiated on two levels, as part of a standard induction plan with newcomers, and as a re-introduction to other employees every six months, to keep them updated. "Companies really need to understand that all internet content that enters, circulates and leaves the building carries a risk and that better understanding at the workforce level is excellent front line defence."IDENTITY THEFT
E-mail identity theft is recognised as a major problem on the internet. According to a 2000 report from the US National Fraud Center, the internet provides the identity thief with the shelter of its anonymity and the speed of its electronic transmissions. It describes the potential harm caused by an identity thief using the internet as 'exponential', and says that incidences of identity fraud are expanding as fast as internet usage. The report, Identity Theft: Authentication as a Solution, advocates improved verification measures to combat criminal activity.
In addition to perpetrating a crime, there have been cases where identity theft has been used by unscrupulous businesses to send 'spam' e-mails to promote their products. And some computer viruses can also execute a form of identity theft. For example, the Klez virus copies the e-mail identities stored in an infected computer user's programs and spreads by sending itself to them. Organisations that suffered the effects of Klez last year included the US State Department.
