Salesforce AI Research paper introduces Just-in-Time agent memory
The shift from write-at-task-end to retrieve-then-curate changes how agent memory pipelines should be designed, favoring raw trace storage plus task-specific extraction. Builders working on long-lived agents can apply the approach to improve success rates on benchmarks like tau2-bench. Compared with traditional summarization-based memory, even an untrained curator already performs competitively.
Salesforce AI Research published a paper on agent memory titled Just-in-Time Memory.
The method stores raw trajectories and defers extraction until the next task arrives, instead of summarizing each run when it ends.
A curator reads retrieved traces together with the new task and writes a short memory payload for that task, and can be trained on whether that same task succeeds.
On ALFWorld, WebShop and tau2-bench, it beats the strongest baseline by 16.2, 16.
3 and 3.9 success-rate points, and even the untrained curator matches or beats task-end memory.
Banger paper from Salesforce AI Research on agent memory.
The overall finding is that you want to store raw trajectories and decide what to extract from them when the next task arrives, instead of summarizing each run when it ends.
Just-in-Time Memory uses a curator that reads the retrieved traces together with the new task and writes a short memory payload for that task. Because the payload is used right away, the curator can be trained on whether that same task succeeds.
On ALFWorld, WebShop and tau2-bench it beats the strongest baseline by 16.2, 16.3 and 3.9 success-rate points. Even the untrained curator matches or beats memory that is written when a task ends.
Paper: https://t.co/0TG98s24dt
This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Salesforce AI Research, agents, paper.
Source attribution: curated from DAIR.AI (@dair_ai).
