Look at your current info-system and for the occurrence of anything that contains natural language information. This includes things like: e-mail, personnel records, research reports, memos, faxes, product descriptions, and word processing documents. Now imagine being able to collect and perform queries against these items as if they were a traditional database.
If you look closely enough you'll probably discover that about half the information that is resident within the organization has natural language or text as one of its most significant attributes. And chances are that there is little or no ability to manage and query these information resources based on their content. Traditional databases are fine as long as you are just adding up numbers or manipulating inventories, but people use language to communicate and no product except Texis can provide real access to the "natural language" components.