(FORTUNE Magazine) – Coldplay, my favorite band, is performing on a Saturday Night Live rerun tonight, but I wouldn't have known about it had I not gotten an e-mail this morning. The message came from something on AOL called MyBestBets, which is powered by a startup called ChoiceStream. Its technology could signal the future of how we find, pick, and learn about what entertains us.

Sorting our way through proliferating entertainment options is a growing challenge. Bruce Springsteen sang about "57 channels and nothin' on" back in 1992. Now the average American household gets about 100 channels of TV and often has access to hundreds more video-on-demand movies and shows. Says Josh Bernoff, who studies entertainment at Forrester Research: "The cable operator almost has a responsibility to give you a tool to better figure out what to watch, now that it's given you all those choices." The TV industry for years has talked casually about the "least objectionable program," the one you end up watching because it's simply not as bad as those on other channels. The challenge is equally big--and getting bigger--in music: With online services like Apple's iTunes Music Store, customers are confronted with a nearly limitless choice.

Now it's up to the tech industry to devise ways to help you figure out what you want. The best known of the personalization engines is at Amazon.com. The online store watches what you buy and suggests other products you might be interested in by matching your purchases to purchase patterns of other shoppers--a system based largely on something called collaborative filtering. I've found plenty of relevant stuff I wouldn't have otherwise known about. TiVo, too, uses filtering algorithms. But all those filters have serious limitations. Buy a baby book for a friend, and Amazon may assume you are a parent. The Wall Street Journal published a story last year on people who are not gay but whose TiVos think they are because of one or two shows they recorded.

Instead of collaborative filtering, ChoiceStream uses more complex approaches to statistical analysis called Bayesian-and discrete-choice modeling. When applied to movies, the technology factors in a variety of data, including professional and user reviews, along with keywords like foreign, Hollywood, romantic, etc., and actors' names. With this knowledge about the content, ChoiceStream can infer quite a bit about your likes and dislikes when you merely pick or pan one film. It also examines the preference patterns of others who have said they like or dislike this movie. And it works. Says ChoiceStream CEO Steve Johnson: "People don't have a lot of patience for crummy results that don't seem tailored to them."

To get started using this technology, you have to spend a few minutes telling it about yourself. I went to AOL keyword mybestbets (or try this out at www.mybestbets.tv) and filled out some forms, listing the few TV shows I like, saying I prefer independent films, then selecting from a list of movies I had seen and rating each one. I told the software how often I watch certain genres like horror (never) and foreign language (often). I then pressed a button, and it gave me a surprisingly useful rundown of shows and movies on my cable system that day, in addition to potentially interesting movies in theaters. I can fine-tune the recommendations, and the software watches my clicks as I navigate the ChoiceStream lists. It gets more accurate over time.

When AOL (which, like FORTUNE's publisher, is a division of AOL Time Warner) releases a new version of its client software at the end of July, the capability will show up in areas for music, shopping, sports, and concerts. This is a breakthrough: AOL has resisted new technology for far too long. Now it is showing signs it has a plan to stanch the continuing loss of members who see little value for AOL in an era of broadband Net access. "We want to make the service more relevant to our members," says AOL personalization vice president Peter Negulescu.

You can also bet that this or similar technology will appear in services all over: on the Internet at MSN or Yahoo, on TV set-top boxes, in the TiVo service, and in some of the many online music-purchasing services on the horizon. "Personalization is the only effective solution to the information-overload problem," says Pattie Maes, an MIT professor and personalization pioneer who sits on ChoiceStream's advisory board. Such technology will be most useful when it's tied automatically to storage connected to our computers and TVs. Then we can come home every day to preselected entertainment we actually like. Meanwhile, I'm studying my daily e-mail. I've never heard of a movie called Manito, but I might check it out later--the e-mail says it's playing four blocks from my house.

David Kirkpatrick is senior editor for Internet and technology. Fast Forward appears weekly on fortune.com and is also available by e-mail subscription.