aiqbench analysis
A new study comparing matched sessions reveals that autonomous AI agents execute approximately 26 minutes of work per session, a stark contrast to the 33 seconds achieved by traditional search assistants. This shift indicates that agents are economically viable for complex, multi-step workflows while remaining complementary to, rather than substitutive of, basic search queries. The findings suggest a clear cost-structure transition where agents become more efficient than conversational modes once tasks exceed a specific step count threshold.
A New Study from Harvard and Perplexity Finds AI Agents Perform 26 Minutes of Autonomous Work per Session vs 33 Seconds for Search
A new Harvard and Perplexity paper uses matched-pair sessions to compare an autonomous agent with a search assistant. It finds large gains in autonomy, time, and cost, plus broader scope of work attempted. The post A New Study from Harvard and Perplexity Finds AI Agents Perform 26 Minutes of Autonomous Work per Session vs 33 Seconds for Search appeared first on MarkTechPost.
A new working research from Perplexity and Harvard offers field evidence on what AI agents do to knowledge work. It draws on production data from two Perplexity products: Search and Computer. The setup is a natural comparison. Search is a conversational answer engine. Computer is an agent that plans and executes tasks end to end. The same users touch both products, so the team can hold the task roughly constant. The research study covers a 90-day window, February 27 through May 27, 2026. Computer launched two days before that window opened. The core method matches near-identical query pairs across the two products. The research team found 10,000 session pairs with cosine similarity above 0.99. Each pair…
Excerpt shown under the publisher's RSS terms · 116 words · full text and figures are on the publisher's site
Related coverage
Humanity’s Last Invention — Richard Socher of Recursive
Latent Space · Sep 14, 2026
The J-Space Debate, Agent Swarms, and Pacing Frontier AI - Digital Minds Newsletter #4
LessWrong · Sep 18, 2026
How AI Agents Reshape Knowledge Work: Autonomy, Efficiency, and Scope
arXiv · Jun 12, 2026
How AI Agents Reshape Knowledge Work: Autonomy, Efficiency, and Scope
arXiv · Jun 8, 2026