The Critical Role of Trusted Agents

About ten years ago I was working with the United Nations, trying to determine why one-third of the banks in a particular country had gone bankrupt. I found that the country’s government did not utilize the concept of either a social security number or an income tax number. Accordingly, individuals could borrow funds from a number of banks without any bank being aware of the other loans. When the creation of a nationwide financial information infrastructure was proposed, bankers eagerly embraced the notion that they could be informed about other banks’ transactions. “That is exactly the system we need,” asserted one bank president. When I informed him that he would also have to share his bank’s information, he balked: “Never. No information from my bank will ever go out to other banks.” The banker’s reaction is not as odd as it sounds. There is a high degree of distrust that exists across all kinds of interorganizational and intergovernmental borders, and it is a major bottleneck to productivity and efficiency.

The first quantum leap in productivity that can be attributed to information technology occurred in the 1960s and ’70s, when automated techniques in factories helped to increase the productivity of blue-collar workers. A similar boost occurred in the 1980s and ’90s, when advances in office computers and related technologies enhanced the productivity of white-collar workers. Now, the challenge is to develop pragmatic ways to increase the productivity of interorganizational tasks and processes, which will provide a third, and probably the most profound, surge in international trade and global growth.

As information-sharing technology has proliferated dramatically over recent decades, organizations increasingly operate on the assumption that any piece of information they release to one particular entity will quickly find its way into the hands of many others. Current technologies enable massive amounts of data to be disseminated instantaneously without the original owner becoming aware of it. Thus, there is a growing reluctance among companies to allow data to be exported. As each organization attempts to maximize its own profits, it fosters local rather than global optimization of industries or industry segments.

Take, for example, the transportation and distribution sector in the United States. The cost of manufactured goods can be broken into various components: the portion attributable to blue-collar labor has decreased significantly over the last 25 years; the portion attributable to white-collar labor has also declined. However, the component attributable to transportation and distribution has actually risen. That is a manifestation of an interorganizational process in which improvements and efficiencies can be brought about only through mutual cooperation and trust, or through incorporation of new paradigms that encourage sharing of information without risking proprietary information or diluting competitive ability.

When two organizations needed to share proprietary data in the past, attorneys, accountants and managers from each would spend weeks or months formulating the details of the relationship. That paradigm is obviously not scalable for use in a networked environment in which the relationships need to be formulated on a near real-time basis, and in which there is a broad and diverse base of potential partners. The problem grows exponentially with the size of the group.

The solution will be the proliferation of trusted agents: organizations, individuals or specialized computer programs that act as information filters. These trusted agents are able to access all data from a set of participating organizations, but chartered to disseminate only parts of that data on a need-to-know basis depending upon the identity of the recipient organization. Different subsets of the data are provided based on whether the recipient is a current customer, a potential customer, a competitor, a supplier and so forth.

The notion of a trusted agent is similar to that of a credit bureau. When a customer applied for a loan 50 years ago in the United States, the bank took down all the details. Today, nationwide credit bureaus collect financial information about individuals on an ongoing basis, and part of that information is released to banks when they need to know. Whereas the credit bureau can collect many types of information, its dissemination is governed by national and state laws, which in turn reflect the negotiated consensus between financial organizations and consumer groups. Through laws and other means, the quantity and the quality of the distributed data can be controlled, as can the timing of the distribution.

Let’s apply that paradigm to the global trade environment. Increasingly, manufacturing companies rely on contractors and subcontractors located around the world. An automobile manufacturer, in order to ensure that its assembly line is not held up because of its computer supplier’s lack of inventory at any given juncture, may insist that it be given full access rights to the supplier’s inventory database. That company may agree to do so if it is a captive supplier to the automaker, but if it is supplying to both GM and Ford, then it of course cannot give full access rights to either. A trusted agent, however, could look at the entire inventory database, and then send “checked, positive” or “checked, negative” messages to GM or Ford as appropriate.

Similar concepts are likely to emerge in many other industries. In the pharmaceutical sector, for example, there are big manufacturing companies (for instance, Merck and Bristol-Myers Squibb), large national and regional distributors (McKesson and Cardinal Health), large decentralized retailers (CVS and Walgreens), as well as independently owned pharmacies. Each retailer manages its own inventory and decides upon the quantity of each drug to keep at each retail outlet and warehouse. That leads to significant duplication of inventory across different chains, and significant waste, since many drug items have expiration dates. Sharing data on a trusted-agent basis enables companies to design inventory systems that reduce aggregate inventory levels. In one scenario using both trusted agents and datamining techniques, inventory was reduced by 50% (down from $1 billion) without adversely affecting the probability of finding an item in stock. Increased supply chain efficiency is one of the main benefits of the trusted agent paradigm. When organizations do not trust one another, they tend to stockpile resources within their respective spheres of control. Once trust is established, each organization opts for much leaner inventory levels.

Some industries, like the financial sector, already have the trusted-agent model, but will likely refine the approach. For example, the Federal Reserve Board currently plays the role of trusted agent in processing checks in the United States, laboriously routing them to appropriate banks nationwide on a “need-to-receive” basis. The total cost for the full cycle — from the point when a check is written to the point when it is returned to the account holder — is estimated at more than $1.25 per check, or more than $62.5 billion for the 50 billion checks processed per year in the United States alone. A new infrastructure in which all checks are sent electronically to an intermediate organization, which then forwards the relevant information in electronic form to the concerned bank or banks, would drastically reduce aggregate costs and processing delays.

In some cases, trade associations will play the role of the trusted agent, just as today the International Air Transport Association serves as the clearing agency for airline tickets. In other cases, auditing companies may play the role, such as in situations where the net profits or total costs have to be distributed among several companies based on data that each company deems proprietary. And in still other cases, government agencies may fill the function; one example could be the selective distribution of personal data relating to individuals when parallel legal claims are submitted to multiple insurance companies, or when addressing emergency situations like the 9/11 tragedy. The value added by the trusted agent could be so high that it could charge for its services. In most instances, however, it may need to be established as a nonprofit entity for it to command the respect of its customer base.

The evolution of trusted agents for different domains will be the key to the next major surge in interorganizational productivity. For collaborating organizations, the instantaneous access to accurate, trusted, timely and appropriate external information will, in turn, create internal efficiencies in manufacturing and planning. Over time, a trusted agent in one domain may interact directly with those in other domains. Ultimately, a web of trusted agents will enable end-users to deal with many different disciplines with a sophistication that exceeds that of individual specialists today.