iMailLight is a Microsoft Outlook^® plug-in and - just like any other software tool - can easily be installed on a desktop PC.

iMailLight offers users a highly-intelligent personal e-mail management assistant that analyzes incoming e-mails and sorts them into different Microsoft Outlook^® folders according to topic and content.

The Microsoft Outlook^® rule assistant is designed to provide similar functionality: however, the rules are far less flexible, requiring a high degree of manual configuration based on a series of individual keywords.

iMailLight on the other hand uses AI (artificial intelligence) methods to analyze incoming e-mails, enabling a significantly more intelligent e-mail sorting process.

A typical e-mail user sets up between 10 to 20 sub-folders to separate messages from friends from eBay transaction e-mails, order confirmations, newsletters and SPAM.

Because iMailLight is a self-learning AI system, its training starts from the initial installation phase using existing e-mail folders and messages. New e-mails received after this phase are automatically transferred to the relevant target folder.

And because the filter function is activated automatically when a new e-mail is received, it is a simple process for users to set up personalized SPAM filters using iMailLight: simply set up an Outlook "SPAM" folder, transfer 20 SPAM messages to this folder and iMailLight will immediately start learning. And your work is done!
In future, all similar e-mails (i.e. SPAM mail) will automatically be transferred to this folder.

The great thing is that you can now add new folders and use iMailLight to sort relevant e-mails for these folders, too. For example, you can harness the tool to transfer incoming eBay-related correspondence to a special eBay folder.

The statistical classification process used by iMailLight can be applied for 10 to 20 categories (folders); the ability to differentiate accurately between categories decreases exponentially after this upper folder limit has been exceeded. Unlike rule-based systems, statistical systems cannot be fine-tuned.

For this reason, statistical e-mail management systems are only useful in environments where the aim is to sort incoming e-mails into a limited number of categories dealing with significantly different topics and message content.