If you are an admin, architect, developer or IT leader attending dreamforce this year, there is a fantastic Salesforce Security session led by Rachel Beard and team on how to build trusted, secure, and compliant AI on the platform. This post highlights what admins and architects will actually get in their 1-hour workshop and companion session— with practical takeaways you can bring back to your org.
What This Workshop Covers
- Einstein Trust Layer, demystified: How it governs AI interactions and enforces “hard rails” for Agentforce.
- Prompt & data protections: Securing prompts, dynamic grounding, data masking, and zero-data retention policies.
- Hands-on with security products: Work in a Salesforce trial org enabled with Shield and Privacy Center.
- Audit-friendly exercises: Walk away with configurations and an audit trail you can show leadership.
Why It Matters for Security
Many teams view the Einstein Trust Layer as a “black box.” Rachel’s session opens it up—showing exactly what is controlled, where policies are enforced, and what admins/architects can configure to prevent leakage from unstructured data and AI prompts. If your CISO asks “How is our data protected with Agentforce?”, this workshop gives you clear answers.
Hands-On Format (No Prereqs)
- Trial org provided with security products focused on AI and the Trust Layer.
- 30 days of access post-Dreamforce to finish exercises and run demos for leadership.
- Experts on standby during the workshop for Q&A.
- Materials provided in print and as links (session itself will not be recorded).
Key Capabilities You’ll Practice
Prompt templates & policies – Standardize and govern prompts to reduce leakage and drift.
Secure data retrieval & dynamic grounding – Attach just-enough context safely to improve responses.
Injection & content safety checks – Detect prompt/response toxicity and multimodal risks.
Sensitive data protection – Enhanced masking, zero-retention, and network protections.
Monitoring, quality, and audit trail – Track usage, quality metrics, and feedback for governance.
Bonus: Tools to Explore After the Session
- Shield – platform encryption, event monitoring.
- Privacy Center – enterprise privacy operations.
- Security Center Data Classification – field-level classification tooling (requires the Security Center add-on).
Beyond Dreamforce: Trusted Agents Workshops
The team is also lining up Trusted Agents workshops in Atlanta and the DC area, with New York and Chicago under consideration. If you are around DC area, there is a great workshop on 9/30 which you should consider attending by clicking on this link.
Who Should Attend
- Admins who need practical guardrails and audit artifacts.
- Architects who must translate CISO concerns into enforceable controls.
- Leaders who want a clear, demo-ready view of AI risk mitigation inside Salesforce.
How to Prepare for the Session
- Bring a short list of objects and fields that you consider sensitive (e.g., PII, PHI, PCI, etc.).
- Note where unstructured data is currently used (cases, notes, files, and email-to-case).
- Decide what you’ll show leadership after the event (masking policy, audit report, or a Trust Layer diagram).
If you have any further questions, feel free to post your comments and email me at buyan@eigenx.com .