Skip to main content

GPTfy + Salesforce Trust Layer for Secure AI Deployments

· 5 min read
GPTfy Team
GPTfy Team
GPTfy Product Team

Illustrates the seamless integration between GPTfy and the Salesforce Trust Layer, highlighting the robust security measures in place for AI deployments within Salesforce. A sophisticated solution that showcases the layers of security, data privacy protocols, and the collaborative framework that GPTfy and Salesforce technologies employ to ensure the highest standards of data protection. This guide emphasizes the commitment to safeguarding sensitive information, mitigating risks, and building trust in AI-powered applications, making it clear that GPTfy and Salesforce prioritize secure, reliable, and ethical AI solutions.

Table of Contents​

55% of reps in any organization use generative AI without permission, making your organization vulnerable to possible data breaches.

This is why having a Layer, aka Salesforce Trust Layer or GPTfy’s Security Layer, is a must to ensure your users are using Gen AI securely.

What?​

An overview of the integration of generative AI within Salesforce, focusing on the security and trust mechanisms provided by the Einstein Trust Layer and GPTfy’s Security Layer.

Why?​

  • To navigate the complexities of integrating generative AI into Salesforce without compromising data security and privacy.
  • To ensure a trusted and compliant AI-driven experience within Salesforce applications.

What can you do with it?​

Use Generative AI securely in your Salesforce without any worries about data breaches.

How?​

Salesforce AI Security Lifecycle, including steps for data extraction, masking, secure retrieval

alt text

Lifecycle for Salesforce, AI, and Security​

  • Extract data
  • Anonymize PII
  • Add context
  • Sanitize content
  • Process prompts
  • Data retention
  • De-anonymize and automate responses
  • Continual monitoring

How Salesforce Trust Layer Secures Data in AI​

alt text

Salesforce AI Trust Layer​

The Trust Layer is a secure intermediary for user interactions with LLMs, ensuring data privacy, preventing user data persistence, and standardizing model interactions.

  • Securing Prompts: Sanitizes and secures all prompts through the Trust Layer before AI interaction.
  • Secure Data Retrieval: Grounds prompts with relevant data to enhance accuracy.
  • Dynamic Grounding: Incorporates additional data and logic for richer insights.
  • Data Masking: Applies entity detection and substitution to protect sensitive information.
  • Prompt Defense: Utilizes defenses to guide model output.
  • LLM Gateway: Ensures secure, compliant AI provider interactions.
  • Response Processing: Performs toxicity detection to ensure suitability.
  • Data Demasking: Reintegrates masked PII securely.
  • Feedback Mechanism: Users can provide acceptance, modification, or rejection feedback on AI quality and accuracy.

How GPTfy’s Security Layer Secures Data in AI​

alt text

GPTfy introduces a sophisticated security layer that seamlessly integrates with Salesforce, offering an advanced mechanism for anonymizing sensitive data before it interacts with AI.

GPTfy Security Layer Process Flow​

  • Secure Data Extraction:

alt text

Dynamically extract permitted, relevant data based on permissions and configurations.
Keep all extracted data securely within Salesforce.

alt text

  • Multi-layered Data Anonymization/Masking

    • Layer 1: Mask sensitive field values using Regex

alt text

  • Layer 2: Anonymize common data types across text fields

alt text

  • Layer 3: Allow blocking of specified sensitive values

alt text

  • Data Security Grounding

    • Process and store data exclusively within Salesforce org
    • Eliminate external third-party data stores
    • Prevent unauthorized external connections
    • Leverage Salesforce user permissions for access controls
    • Encrypt data in transit between internal components
    • Harness Salesforce Platform Shield for security monitoring and threat detection

  • Zero Data Retention in AI/LLM

alt text

  • Transiently process data with no retention post-processing

  • Maintain data privacy regulations and Salesforce data integrity

  • De-Masking/Anonymization

alt text

  • De-Mask PII sent to AI

  • Re-identify anonymized data so you don’t have to figure out the sensitive data removed by GPTfy

  • Automate

    • Update the fields (e.g., Case.sentiment -> Positive)

  • Feedback

alt text

  • Users can give thumbs up/down for AI responses
  • Feedback helps prompt engineers iterate and improve prompt outcomes

Key Takeaways​

  • Multi-layered security by GPTfy and Salesforce provides a robust solution for anonymizing sensitive data in Salesforce.
  • Configuration allows for tailored anonymization strategies across data types and fields.
  • Security audits showcase the effectiveness of the implemented security layers in protecting PII.

Benefits​

Leveraging GPTfy’s multi-layered security features ensures compliance with data protection regulations and fortifies Salesforce environments against data breaches, providing peace of mind for businesses and their customers.

Conclusion​

The GPTfy’s security layer represents significant advancements in securing generative AI within Salesforce, providing frameworks that protect data and ensure that AI-driven processes are compliant and trusted.

By leveraging these capabilities, Salesforce developers can confidently integrate AI into their Salesforce org, knowing that their sensitive data is secure and their interactions are safe.

Explore further how GPTfy can transform your Salesforce AI applications by ensuring security and trust at every step.