Perplexity, in collaboration with the Harvard Business School, has released a new research report that intuitively demonstrates the comprehensive transformation AI agents bring to knowledge work through a comparison between Perplexity Computer's general AI agent orchestrator and traditional search assistants. Traditional AI assistants can only answer questions, with subsequent operations still requiring manual user completion. In contrast, AI agents can autonomously plan tasks, execute them, and output final results.
Research data shows that under the same task scenarios, Perplexity Computer's AI agent runs autonomously for an average of 26 minutes per session, which is 48 times longer than the 33 seconds of traditional search assistants. This agent frequently calls various tools across platforms, significantly enhancing its automation capabilities. With its strong autonomy, it performs exceptionally well in terms of work efficiency and cost control. Compared to traditional models, task completion time can be reduced by 79% to 92%, and overall costs can be cut by 87% to 96%, with the most significant optimization effects seen in programming.
At the same time, AI agents have also expanded the boundaries of work. Users no longer need to perform basic operations one by one, and their roles gradually shift from frontline operators to work supervisors, capable of handling more complex and broader tasks. The report suggests that in the long term, AI agents will further impact job classification, personnel allocation, and team structure, continuously reshaping the nature of workplace tasks.
Key Points:
🤖 There are clear differences between AI agents and traditional AI assistants; the former can complete tasks throughout the entire process autonomously, with significantly longer autonomous operation times than the latter.
⚡ AI agents greatly optimize work efficiency and cost, achieving significant reductions in task duration and overall costs, benefiting all fields broadly.
👥 AI changes the user's work role, expanding the scope of work. In the long run, it may reshape job definitions and team structures.
