• enables individually directed information management
• complexity
• volume
• alert users to what is significant.
• agents can take action - if so directed.
• the Professor Pattie Maes (MIT MediaLabs) theory: beyond 'foraging'

What's the goal of General Magic in building Agentware? I think there are a common set of themes which I have attempted to capture on this slide. (There are, as Tony Rutkowski would suggest, some common types of agents which seem to be evolving)

Alert Users To What is Significant...Enable timely, knowledgeable decisions & actions

Advisory: suggest. monitor. These are mostly static and local to the user (though they need not be)

Assistant: messaging, e-commerce (these are the butlers of Cyberspace, andt hey are likely to be mobile and commerce aware)

Information: filtering, alerting, directing, inquiring.

Society must evolve from gatherers & shufflers to survive the new age in which we find ourselves...Most of our interactions which focus on turning data into information or information into knowledge are too much like random forging, much effort yields little return. This is because the tools we use to-day were never designed for such tasks; Agentware is. Let's look at what I mean by this...

Agentware, a presentation by Dana Moore of AT&T.
dana.moore@worldnet.att.net