To make an Agentforce agent remember useful context, first decide what needs to persist and for whom. Salesforce offers several different mechanisms: conversation-derived memories for one user and one agent, Data 360 memory and context for continuity across agents, variables for workflow state, profile fields for user attributes, and Data Library retrieval for organizational knowledge. They are not interchangeable, and only some are cross-session memory.
What “memory” means in Agentforce
Persistent memory is not one shared store that automatically carries every detail between conversations. The right mechanism depends on the scope: a value needed by the next action, a detail an agent should recall for the same user later, or context that should follow a person between agents. Separate from all of those is company knowledge the agent retrieves to ground an answer.
As an Amazon Associate I earn from qualifying purchases.
The distinctions below reflect Salesforce Help and Salesforce Developers documentation reviewed on October 7, 2026. They describe vendor-documented capabilities, not independent tests of memory accuracy, response quality, latency, or cost.
| Mechanism | Scope and source | User control and governance | Data 360 dependency |
|---|---|---|---|
| Agent Memory | Conversation-derived details for one user-agent pair across conversations. | Employee and Service agent behavior differs; see the consent and management section. | Not stated as a requirement for Agent Memory on the reviewed Help page. Check current entitlements with your Salesforce administrator. |
| Data 360 Agentic Memory and Context | Persistent conversation memory and extracted facts, preferences, and summaries; can support continuity across agents when linked to a Unified Individual. | Salesforce says object-, field-, and record-level access controls are respected. | Data 360 is central to this documented cross-agent context capability; verify the target org’s setup and licensing. |
| Agentforce variables | Structured values passed into or out of actions and used to control action or topic execution. | Workflow state is explicitly represented in the agent’s action flow. | Not stated as a requirement in the cited Salesforce Developers article. |
| Profile context | Record fields such as name, locale, or time zone. | Profile data is separate from the memories a user views or manages. | Not stated as a requirement on the reviewed Help page. |
| Data Library retrieval (RAG) | Retrieved organizational knowledge, such as indexed files, fields, or supported web sources. | Retrieval depends on configured sources, permissions, and retriever behavior. | Requires Data 360 setup and consumes Data 360 credits, according to Salesforce. |
When to use Agent Memory for one user and one agent
Use Agent Memory when an agent should personalize later conversations using relevant details identified in earlier conversations with the same user. Salesforce describes it as identifying relevant conversation details, storing them, and using them as context in future conversations.
Memories are maintained separately for each user-agent pair; one agent’s memories are not automatically available to another agent. Salesforce documents a maximum of 50 memories per pair. When that limit is reached, the oldest memory is removed. This is a limit on stored memories, not a stated transcript-history or token limit.
Memory can be disabled without deleting the stored items: Salesforce says the agent stops using them while disabled, and can use them again if memory is re-enabled. Plan explicitly for what users should be able to inspect, correct, or remove, rather than treating saved conversation facts as invisible application state.
How consent and memory management differ by agent type
Employee agents
For Employee agents in Lightning Experience, Salesforce says capture begins only after the user opts in. Make the opt-in behavior clear to users and confirm the experience on the surface where the agent will be used.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsService agents
For Service agents, users can manage memories conversationally if the builder adds the User Memory Management subagent. Salesforce says this subagent is not added automatically. The documented Service-channel experience does not have a separate disclaimer or opt-in step, so communicate collection and management behavior through the service experience and your organization’s policies.
Salesforce’s example of a memory-management request is “Show what you remember about me.” Confirm the supported management interaction in the target agent before launch; do not assume that a management subagent is present just because Agent Memory is enabled.
When Data 360 memory and context are the better fit
Choose Data 360 Agentic Memory and Context when the requirement is persistent context that can support handoffs or continuity across multiple agents, rather than isolated recall by one agent. Salesforce documents extracting facts, preferences, and summaries on a periodic cadence, retrieving context on demand through GetContext, and linking memory to a Unified Individual.
Salesforce says conversations, session summaries, and profile memories become available for context retrieval “within seconds of ingestion.” That is a product capability statement, not an independently measured timing guarantee or a universal service-level commitment. Build and monitor the experience for the freshness your use case actually requires.
Documented scenarios include multi-turn support, sales follow-up, web-agent personalization, handoffs, and employee-assistant experiences. Because this design ties context to a Unified Individual and uses Data 360, validate identity matching, access boundaries, and entitlements before depending on cross-agent continuity.
Use variables for workflow state, not long-term recall
Variables are the better fit when an agent must carry a defined value through an action sequence—for example, retaining troubleshooting steps as a conversation progresses or deciding whether a later action should run. Salesforce Developers describes variables as “a type of structured memory whose value can be used as an input or output for actions.”
This is short-term, structured workflow state, not a substitute for cross-session user memory. A variable can represent a known value in the flow; Agent Memory instead stores conversation-derived details for later interactions. Use explicit variables when predictable action sequencing matters, and use a persistent memory mechanism when the context must survive beyond that workflow.
Use profile context for stable user attributes
Fields such as a user’s name, locale, or time zone can personalize an interaction without being saved as conversation memories. Salesforce says profile context works independently from Agent Memory and does not appear in the memories users view or manage.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPrefer a profile field when the information belongs in a governed user record and should be maintained as profile data. Do not rely on memory extraction to keep a canonical attribute accurate when the system of record can provide it directly.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Use Data Library retrieval for company knowledge
Agentforce Data Library indexes organizational material—such as knowledge, fields, uploaded files, or supported web sources—so an agent can retrieve information to ground an answer. This is retrieval-augmented generation (RAG): it supplies relevant source material to the response, rather than proving that the agent remembers a particular user’s earlier conversation.
Salesforce says Data Library requires Data 360 setup and consumes Data 360 credits. In Data Cloud One companion orgs, Salesforce says Knowledge and Files are not supported as Data Library sources and recommends a custom retriever. Check the supported source types for your org, configure retriever permissions, and account for credit use before selecting this architecture.
Choose a design by scope, control, and dependencies
- One user, one agent, later conversations: consider Agent Memory, then confirm the relevant agent type, consent behavior, and user-management experience.
- One workflow, several actions: use variables for the structured values and conditions that need to pass through that workflow.
- Continuity across agents: evaluate Data 360 Agentic Memory and Context, including Unified Individual linkage, identity handling, and access controls.
- Current user attributes: supply profile context from the appropriate record rather than expecting a conversation memory to act as a system of record.
- Answers grounded in company material: configure Data Library retrieval or another supported retriever; do not label retrieved knowledge as personal memory.
For each candidate, decide what data it can read, who can change or delete it, how fresh it must be, what happens when it is unavailable, and whether it should be isolated to one agent or shared through an identity-linked context layer. Salesforce documentation explains the feature boundaries but does not establish independent performance comparisons or a universal cost comparison.
Recommended Free Tools
Check product status and org readiness before implementation
Salesforce’s setup page says Agentforce (Default) stopped receiving new features or improvements and was no longer available in new Salesforce environments beginning June 17, 2025; Salesforce recommends migration to Agentforce Employee agent for continued enhancements and support. Edition availability and add-on licensing vary by agent type. Confirm the current product labels, entitlements, and migration plan for the specific org rather than following an older guide as if every environment had the same setup.
Quick Recap
- Confirm which agent type and channel are in scope, and whether Agent Memory is supported there.
- For Employee agents in Lightning Experience, define how users opt in; for Service agents, decide whether to add the User Memory Management subagent.
- For cross-agent context or Data Library, confirm Data 360 configuration, required licenses, source support, retriever permissions, and credit implications with the Salesforce administrator.
- Set user-facing expectations for memory collection and management, and test the intended view, update, or deletion paths in the target agent.
- Keep workflow variables, profile fields, personal memories, and retrieved knowledge distinct in the design and in access-control reviews.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
