By Itai Asseo and Egor Pakhomov · Salesforce BlogYou will also see how a policy that treats updates as corrections differs from one that files them as new facts.
You will learn how Salesforce AI Research measured agent memory end to end and that a revised memory policy, not an architecture change, roughly doubled recall in Agentforce.
Why read this
You will also see how a policy that treats updates as corrections differs from one that files them as new facts.
Previous policy
Keeps less of what customers say
Files an update as an extra fact
Updated policy
Keeps more of what customers say
Treats an update as a correction
The updated memory policy compared with the one it replaced
Salesforce AI Research built an evaluation framework that generates long, multi-session conversations with specific facts planted along the way, runs them through the live Agentforce system, and then asks questions that depend on those facts. Testing the full path, from the moment a customer mentions something to the moment an agent acts on it, is presented as what makes the result trustworthy. The conversations come from ConvoMem, a public benchmark released in 2025.
The framework was used to improve Agentforce Memory, which became generally available in September. The updated memory policy holds on to more of what a customer says and treats an update as a correction to something the agent already knows, rather than as an additional fact stored alongside the first. According to the article, the policy recalls 1.8 times as many explicitly stated facts and 2.2 times as many facts that change over time compared with the policy it replaced. The underlying architecture is unchanged.
The authors argue that measuring in production matters because memory only behaves like memory under real conditions, and they cite a travel platform customer whose employee-facing agent now carries context across conversations. The article is largely promotional and gives few methodological details beyond the headline ratios.
By Roopang Chauhan, Sundar Vedula and Peng-Wen Chen · Salesforce EngineeringReaders will understand why memory was moved into the Reasoner and how to enable it using Agentscript.
You will learn how the Agentforce team designed persistent memory to solve stateless agent limitations, including retrieval strategies, latency tradeoffs, and privacy controls.
Why read this
Readers will understand why memory was moved into the Reasoner and how to enable it using Agentscript.
The Salesforce Engineering team built Agentforce Memory to eliminate the stateless nature of enterprise AI agents. Without persistent memory, users must repeatedly provide context, which increases friction and slows task completion. The team standardized industry terminology around memory types before designing the architecture.
Early prototypes relied on recent conversation history, but the team shifted to semantic retrieval to capture older but relevant information. They evaluated multiple approaches and selected keyword search during pre-orchestration to keep the context window bounded and minimize latency. Moving memory directly into the Reasoner replaced manual action configurations, simplifying administration and improving developer ergonomics.
Privacy and user control were treated as core architectural requirements rather than afterthoughts. Administrators can enable memory globally, while individual users manage preferences through a console or conversational commands. Enabling the feature requires adding a context.memory block to Agentscript, allowing the platform to handle curation and injection automatically.