For the past several years I’ve made my living in technical support. While this profession is often maligned, it is crucially necessary. Moreover, it provides a unique window into applied epistemology. Epistemology is the field of philosophy that inquires into the nature of knowledge. Tech support provides a very pure example of the problems facing the human mind in its quest for knowledge at the individual and social levels.
A user contacts tech support because something has gone wrong with their software and they hope that tech support will be able to remedy their problem. The mission of a tech support organization is therefore to organize knowledge of the software and deploy it to be able to solve user problems. Their first point of contact in the tech support organization is the tier 1 support agent. To be able to solve a variety of problems at a high enough rate that an insurmountable backlog doesn’t pile up for the organization, the agent must be able to solve a reasonable amount of problems without burdening the organization’s experts. This requires a thorough knowledge of theoretical functioning of the software, but also a great deal of practical knowledge—the knowledge needed to produce certain effects, like knowing how to process credit card payments correctly in a system. This type of knowledge can often be found in documentation, which is a formalized way of writing down how a software system is expected to function, and is endorsed by the organization for public usage.
In addition, there is also a certain amount of folk knowledge needed as well. Folk knowledge is a term from anthropology describing bodies of knowledge that belong to particular tribe of people without being mediated by institutional or formal structures. In this case the tribe is the tribe of tech support agents who, while they are technically part of the same institution, constantly communicate their solutions back and forth with each other through informal channels and build up shared bases of knowledge in greater depth than any other users of the product. Examples of folk knowledge in tech support are things like knowing that a setting needs to be set a certain way, even though it is not specified in the public facing documentation.
Empirical inquiry is also needed to expand the boundaries of what is known directly. When the outcome of an operation in the software is not known—or when the observed outcome is different than the desired or expected outcome—empirical testing is needed to verify the ground truth behavior of the software. In this mode of inquiry, the agent tries to experiment to reproduce behavior consistently and find the conditions that are required for a given outcome. Sometimes the problem cannot be reproduced—when testing, the standard method may work works without error. Other times, the directions the user gives directly result in the same error being thrown. Whatever the outcome, in this mode of knowing the tech support agent must confront the behavior of the software system—a microcosm of nature—directly rather than relying only on the codified abstractions of book knowledge. The tier 1 agent uses all the knowledge at their disposal, including formal knowledge, folk knowledge, and empirical knowledge, to diagnose and treat the problem presented to them.
The problem of social epistemology already comes into play for the tier 1 agent, since so much of their knowledge comes through transmission from peers and those higher in the hierarchy. But the specific issue of organizational epistemology comes when the tier 1 agent is not able to solve the problem with their own knowledge. In this case they must rely on tier 2 (or higher tiers, lumped in with tier 2 here) of tech support. The existence of tier 2 reveals that the tech support organization is a bureaucracy—it is a hierarchical organization staffed by experts who are the authoritative voices on the product; further, the tiers of support must interact with each other through formal protocols.
To receive assistance from tier 1, the agent must ask a precise question, provide the customer’s problem in writing, state the software configuration as well as all the troubleshooting steps that were already performed. In other words, the tier 1 agent must reduce their previous work on the case, which was more than likely a sprawling and ad hoc communication with the live customer into a rational, legible question that is easily understood. It is key that tier 2 receives communications that fit this format because of the large number of issues they must deal with; the purpose of the bureaucratic process is to maximize the effectiveness of the scarce resource of their expertise. Tier 2 serves as the authoritative expert on the software product. They must determine whether a given issue has a solution, has no solution, or must be sent to the software development team to solve. For the development team to look at an issue, usually a reproduced version of the problem is required so that they can do the tests necessary to isolate the actual problem in the code. Tier 2 must try to create this reproduced version to send to them.
A great many problems in software can be solved by tech support. There is knowledge contained between documentation and the heads of experts which can turn a painful situation on the user’s end into a perfectly working solution in very little time. But not all problems can be solved; the trickiest problems may not admit any solution. This is similar to how in medicine, sometimes the doctor runs out of options and there is nothing that can be done but let the patient die; expertise does not guarantee answers to every single problem. But as a general rule, we organize our social systems to make the best use of expert knowledge to solve as many problems as can be solved. We experiment and try to find steps through which issues and solutions can be consistently reproduced; we communicate our knowledge to each other orally and informally, or through formalized documentation; we store knowledge in writing and in the heads of experts. These are key epistemological building blocks that we see clearly illustrated in the field of tech support.
Komentar