Science of AI Prompts in Salesforce: The Prompt Lifecycle with GPTfy – Part 2

Table of Contents

Welcome back! Part 1 covered the ‘What’ and ‘How’ of crafting effective prompts.

Part Deux, in the words of Willy Wonka, “Keeps on truckin’. ” Here’s what we have here.

4. Salesforce Trust Layer: Enforce Data Privacy & Security

5. Testing and Refining Prompts

6. Publishing Prompts for Users

7. Tracking Value and Usage


4. Trust Layer: Enforce Data Privacy & Security

In the world of AI, safeguarding sensitive data is a lot harder because we now have more moving parts – and a Gen AI black box called LLM (Large Language Model).

This makes it harder for CISOs and other InfoSec professionals, who are already overburdened by cyberattacks, and other security challenges.


65% of IT leaders experienced a security breach between June 2021 and June 2022.

Some of the key concerns with AI:

  • What data is retained by AI after a session?
  • Is there a potential for inadvertent exposure to other users?
  • Can AI leak data to external agents?
  • How do we comply with Privacy regulations?

What is a Trust Layer? And how can your AI get one?

Imagine placing a protective layer around generative AI, akin to safeguarding your organization. While internal alterations may be limited, the focus shifts to shaping inputs and scrutinizing outputs for enhanced trustworthiness.

A security trust layer is a protective framework implemented to ensure the privacy and security of sensitive information.

Multi-layered Security

It involves measures such as data privacy tools, including customizable block lists, regex transformation, and pattern masking to enforce data rules automatically when processing prompts.
Check out Salesforce + AI Best Security Practices

Exposure to critical information, such as SSNs, passwords, or financial details, can lead to security risks and compliance headaches.

Using a multi-layered masking with block lists & RegEx can prevent senstive data & PII/PHI from ever leaving your Salesforce Org.

  • Block Lists: Proactively exclude high-risk terms, reducing the likelihood of sensitive data exposure.
  • Regular Expressions (Regex): This technique utilizes regex patterns to identify and mask sensitive data, ensuring that it remains confidential.
Multi Security Layers in GPTfy
  • No Third-Party Data Storage: GPTfy doesn’t use any external data storage. All operations are confined to your Salesforce organization or your Salesforce AI instance.
  • Data Masking for AI Security: GPTfy masks sensitive information before it goes to AI. This includes removing sensitive data from fields and unstructured text using regular expressions and blocklists.
  • Re-identification of Data: After AI processing, GPTfy re-identifies the masked data, ensuring no sensitive information is leaked.
Security Audit

5. Testing and Refining Prompts

Refining a prompt means fine-tuning the instructions or questions you give to the AI to get the best and most accurate results.

Think of it like teaching the AI how to understand and respond to your specific needs within the Salesforce environment.

Enter the realm of iterative perfection! Prompt engineering demands continuous testing and refinement.

Create sample data scenarios, review outputs, and tweak until the results align with your defined expected outcomes.

Here’s how you can do it:

  1. Conduct trial and error! Experimenting helps you discover the most effective way to communicate with the AI.
  2. Be specific! The more specific you are, the better the AI can understand and generate relevant responses.
  3. Identify and include keywords related to your Salesforce tasks.
  4. Adjust prompt details and data fields based on test results.
  5. Aim for prompts that meet user expectations effectively.
  6. Provide details about the data, tasks, or processes you’re dealing with. This gives the AI more context to generate accurate responses.
Detailed Prompt in GPTfy to get an overview of opportunities

Optimize prompt performance seamlessly with GPTfy’s agile, self-service workflow. Test, refine, and validate prompts iteratively without involving IT, ensuring your AI-generated results align perfectly with expectations.

Bring Gen AI in your Salesforce

Why spend thousands of dollars when you can get a Free trial of AI in your Salesforce securely and in compliance with regulations?

Get GPTfyWatch Demos

6. Publishing Prompts for Users

Prompt Catalog

Publish prompts to intended user-profiles and tie them to relevant record types. This strategic move ensures that user groups derive maximum value from AI on a daily basis.

GPTfy simplifies the process, allowing admins to effortlessly publish polished prompts to select Salesforce user-profiles and tie them directly to appropriate record types for widespread user access.

7. Tracking Value and Usage

Engaging users is like throwing a party, and your dashboard wants to be the life of it. Regularly tracking usage helps you know if it’s a wild bash or a quiet soirée.

ROI Dashboard

After all, in the world of dashboards, those numbers speak louder than words!

Measure time savings, identify prolific users, and analyze activity trends by prompt, user, and object.

GPTfy offers dashboards that showcase the ROI of prompts through metrics on time savings, user adoption, and overall utilization.

Prompt Usage by Users, Object, and Feedback

Track usage patterns to identify power users and popular prompts and measure ROI in tangible terms for organizational benefit.

Get ready for a hands-on journey!

Stay tuned for Part 3, where you will learn how to map relevant fields, set data rules, test iteratively, publish profiles to Account Managers, and showcase the tool’s flexibility and scalability in a Salesforce environment.

And there you have it – a comprehensive exploration of prompt engineering best practices brought to you by GPTfy’s robust capabilities.

From defining expected outcomes to empowering users with secure and efficient AI, you’re ready to witness the power of AI firsthand.

Let me know what you guys think, and what else I should talk about in part 3.
PS: If you have any questions, feel free to comment those.

Picture of Saurabh Gupta

Saurabh Gupta

Saurabh is an Enterprise Architect and seasoned entrepreneur spearheading a Salesforce security and AI startup, with inventive contributions recognized by a patent.

Blogs you may find interesting