AT&T Wireless must look like a great buyout target--Cingular and just about everyone else is expressing interest in the company. But what will be most interesting is whether any of these proposals go through (or if so, at what unexpected expense) because of dreadful operations issues within the "company America trusts."

AT&T Wireless owns two overlapping networks, GSM (the newer network of the same technology widely used in Europe and Canada) and TDMA (the incumbent network that cannot support Internet applications). AT&T Wireless has been trying to shift as many customers as possible to GSM, because this opens the door to higher margin data services. However, when a customer switches from TDMA to GSM, the account must be re-created in a totally separate database. This migration winds up becoming irritating for the customer or expensive for the company--neither of which is a good situation.

These problems may or may not be related to the other widespread problems with AT&T Wireless's GSM customer database, which malfunctioned to the point where for weeks, the company was having colossal problems enrolling new customers. Wireless number portability rules beginning November 24 only made a bad situation worse. The result of all this has been an operations disaster with the primary symptom of call center hold times exceeding 30 minutes. Currently, one person I know has been waiting five weeks for AT&T Wireless to figure out how to send him a shipping label--so he can return a phone and end his free trial.

As if that's not enough, in a recent customer satisfaction survey, AT&T Wireless placed at or near the bottom in all 12 major U.S. markets. The company's response to these results was that the survey actually misses the point; what is more important is how wireless telephony has "changed people's lives." OK.

So why am I with AT&T Wireless? Their services are a good fit with my needs, and after some haggling, they agreed to give me a break on a multi-band phone that creates an acceptable situation for now. But I'm not considering whether to pay $30 billion for a company. My question would be whether this $30 billion includes the cost of untangling a network of quadruple knots.