02 / IT · Analysis
TTSE separates agent experience into facts and tips
TTSE proposes a way to classify agent experience, while reported gains remain the development team's own results.

openJiuwen unveiled TTSE, a design that separates AI agent experience into environment facts and conditional tips, according to Pandaily. For developers building agents, AI systems that perform tasks, it offers a way to organise retained experience.
The distinction concerns what an agent keeps from its work. Descriptions of the environment serve a different role from advice that applies under particular conditions. Separating them provides a basis for examining how experience is recorded.
Pandaily also reported fewer tool calls in office-task demonstrations. It presented those gains as the development team's own results. The proposed classification and the evidence for its effects therefore warrant separate consideration.
Environment facts as a separate category
According to Pandaily, TTSE treats facts about the environment as one category of experience. It places conditional tips in another category.
The distinction can be read in terms of the roles of facts and advice. Recording a description of an environment is different from retaining a tip for use when certain conditions apply. That difference is the central design choice for developers to assess.
The scope of the team's reported gains
The reduction in tool calls reported by Pandaily came from office-task demonstrations. Its attribution matters: these are results reported by the development team.
The classification scheme and its reported effects raise separate issues for evaluation. For developers, the next step in assessing the proposal is to consider whether the categories are useful and how far the team's demonstration results apply to other tasks.