Smokers (N = 756) were randomly assigned by stage of change to (a) standardized self-help manuals (ALA+ condition), (b) individualized manuals matched to stage (TTT condition), (c) interactive expert-system computer reports plus individualized manuals (ITT condition), or (d) a personalized condition with 4 counselor calls, stage manuals, and computer reports (PITT condition). Over 18 months, the ITT group's results more than doubled those of the ALA+ group on abstinence measures. The ALA+ and TTT conditions were equivalent over 12 months, but at 18 months the TTT condition was more effective. The ITT condition was the best or comparable with the best treatment at all follow-ups for smokers at all stages of change. Results suggest that an effective expert system has been developed, and discussion focuses on delivering this system to entire populations of smokers.