BOT2001 Report
Wireless Bots Face Unique Challenges

By Brian Proffitt

When we were kids, the world was mostly likely brought to our doorstep in the form of National Geographic. Later, the world came to us through television news, which eventually became a 24-hour affair. Sometime after that, the Internet arrived with the promise of delivering even more content.

But the premise of all of these mediums is that we would be sitting down and taking the time to read, watch, or otherwise peruse these sources of information. In reality, we all have things to do, places to go. Who has time to sit and read all of this stuff?

Today, though, the information is starting to catch up to us as we ricochet through our busy lives. Instead of looking for knowledge, the knowledge is looking for us.

With the advent of cellular and handheld device technology, the chance to deliver information to users any time or anywhere is certainly there. But just because a thing can be done does not always mean it will be done. After all, who has time to look for desired content in a little handheld?

This, then, is the realm of the wireless bot: automated apps that will scour the Internet's data sources and pull in the desired data to your wireless device so it's there ready for you to use.

Wireless bots were the topic of the delivery made at BOT2001 by Dr. Michael Pazzani, Professor at the University of California, Irvine. Pazzani is currently on leave from UC and serving as CEO of AdaptiveInfo.com, a company which focuses on bots that deliver personalized information to wireless devices.

Pazzani agrees with this assessment of the perfect world where content is delivered effortlessly to those who need it. But this is an idealist's view. In reality, he told the participants, users must contend with "slow, expensive connections on devices with small screens."

Pazzani sees wireless bot technology as being the bridge to the gap between reality and the vision for wireless content delivery.

At the center of Pazzani's plans for wireless bots is a concept he calls "adaptive personalization." Instead of using traditional methods of polling a reader for their interests in story topics, a wireless bot could implicitly monitor the story choices the reader makes. By contextually analyzing what the user has listened/surfed to in the past, the bot will be able to present more interesting stories to the reader.

Pazzani asserts this is better than using polling methods such as check boxes, which often miss the mark as far as delivering stories, and certainly will miss delivering breaking news on new topics.

Pazzani has the numbers to back this up. He compared 3000 users reading stories from a Palm handheld device, using both stories coming from a static source and stories fed from an adaptive source. In his study, Pazzani found that reader of the stories coming in on the adaptive queue were far more likely to actually read the story in its entirety.

With adaptive personalization, Pazzani explained, content such as news, restaurants, and even classified ads could be instantly updated to the user's preferences and even location.

Still, a lot more effort will be needed to make wireless bots truly useful, as Pazzani sees this technology as just part of a larger solution. Pazzani wants to develop tools that will address the entire user experience, as well as backend tools to track user and traffic stats.

"Everyone in the value chain must make a profit," Pazzani said, to make such technology work.