This is a true story:
A customer support manager in an IT company thinks the existing customer database is too unorganized. Any support issue older than a few months old at best will need to go through a lot of departments like logistics, manufacturing, and IT to collect information, but most often crucial information like hardware configuration, software version will just be anyone's guess.
The support manager announced proudly he would implement some form of data keeping: customers or vendors gotta fill up support requests, while all the support engineers will need to record down all transaction histories with customers, in MS-Word files.
All the MS-Word files will be saved somewhere as part of the support database.
In case you, my dear reader, don't understand how comical this is, allow me to explain.
First and foremost, these files serve _no_ purpose if they are not integrated and accessible by all relevant stakeholders. For example the support engineer has no easy way to tell what configuration a customer has, and the logistics department doesn't know what kind of bulk discount to give as again there is no easy way to tell what a customer has bought from the company.
Second, there is no known reliable and fast method to search multiple Word files. If you have 100 Word files and need to search through them for a particular customer, good luck. Moreover, Word file format is proprietary and it is at M$'s mercy if future MS-Word will be able to read back old Word documents. Even though the latest docx is in XML format, the question still begs for a rational and technically sound answer.
The most straight forward way that I can think of would be implementing a web-form that updates a centralized database. In this way, the issues are much easier to track.
It is so sad to slowly realize Dilbert actually isn't that absurd.