Salesforce

Public documentation review covering Agentforce and the Einstein generative AI trust layer. Findings reflect the linked first-party sources as reviewed on 4 October 2026. Private and unlinked material was not assessed.

Evidence status summary

Found

Found evidence checks

Partial

Partial evidence checks

Not found

Not found evidence checks

Unclear

Unclear evidence checks

Not applicable

Not applicable evidence checks

Source-linked review

Evidence checks

Each row records the public documentation signal, source links, freshness, and thematic framework tags.

Review methodology

Evidence check	Status	Summary	Sources	Last checked	Themes
AI-specific privacy/data handling statement	FoundMedium confidence	The trust layer describes masking, permission-aware grounding and retention controls.	Einstein Trust Layer	Oct 4, 2026	None recorded
AI usage disclosure	FoundMedium confidence	The documentation describes generative AI and Agentforce architecture.	Agentforce trust architecture	Oct 4, 2026	None recorded
Customer data training statement	FoundMedium confidence	Salesforce states that third-party model partners do not train on data processed through the trust layer.	Einstein Trust Layer	Oct 4, 2026	None recorded
Framework references or governance alignment	UnclearLow confidence	The reviewed public sources do not establish this signal for Agentforce and the Einstein generative AI trust layer. This is a limit of the reviewed material, not a finding about private practices.	Einstein Trust Layer Agentforce trust architecture	Oct 4, 2026	None recorded
Human oversight statement	FoundMedium confidence	The developer guide explains controls and auditability around agent operation.	Agentforce trust architecture	Oct 4, 2026	None recorded
AI incident response statement	UnclearLow confidence	The reviewed public sources do not establish this signal for Agentforce and the Einstein generative AI trust layer. This is a limit of the reviewed material, not a finding about private practices.	Einstein Trust Layer Agentforce trust architecture	Oct 4, 2026	None recorded
Model card/system card or equivalent	UnclearLow confidence	The reviewed public sources do not establish this signal for Agentforce and the Einstein generative AI trust layer. This is a limit of the reviewed material, not a finding about private practices.	Einstein Trust Layer Agentforce trust architecture	Oct 4, 2026	None recorded
Model/provider disclosure	FoundMedium confidence	The trust layer documentation identifies external model services including OpenAI.	Einstein Trust Layer	Oct 4, 2026	None recorded
Responsible AI/governance policy	FoundMedium confidence	The guide links Salesforce's generative AI principles.	Agentforce trust architecture	Oct 4, 2026	None recorded
AI safety/testing/evals/red-teaming statement	PartialMedium confidence	The guide describes toxicity detection and safeguards; detailed evaluation results are outside this reviewed set.	Agentforce trust architecture	Oct 4, 2026	None recorded
Security/trust centre availability	UnclearLow confidence	The reviewed public sources do not establish this signal for Agentforce and the Einstein generative AI trust layer. This is a limit of the reviewed material, not a finding about private practices.	Einstein Trust Layer Agentforce trust architecture	Oct 4, 2026	None recorded
Subprocessor or AI vendor disclosure	UnclearLow confidence	The reviewed public sources do not establish this signal for Agentforce and the Einstein generative AI trust layer. This is a limit of the reviewed material, not a finding about private practices.	Einstein Trust Layer Agentforce trust architecture	Oct 4, 2026	None recorded

Source list

Einstein Trust Layer Agentforce trust architecture

Framework themes

No framework tags recorded.

Help improve this profile.

Submit a correction or claim profile review for free. Updates are based on supporting public evidence.

Submit correction Claim profile