Salesforce environments are constantly growing in complexity, and the administrative effort required to maintain and optimize these environments can be enormous. Metazoa’s Snapshot product has been enriched with innovative Artificial Intelligence (AI) capabilities to assist Salesforce administrators, developers, and architects with the challenges of managing large and complex orgs. While Salesforce has made tremendous strides in enabling AI for sales and marketing end users, Metazoa’s AI features focus on complementing administrative and development processes—removing technical debt, documenting complexity, and enhancing security. This expanded whitepaper offers deeper insight into how Snapshot’s AI features can be leveraged for best practices in org management.
AI Security
Before implementing AI features in Snapshot, it is crucial to understand that AI usage within the platform is entirely optional. You can configure your AI usage in the Snapshot Preferences panel by disabling AI integrations if your organization is still determining its AI policies and practices. Importantly, all core functions in Snapshot remain available, even if AI is disabled.
There are ongoing concerns around data privacy and security, and we share those concerns. To address this, Snapshot communicates directly with the OpenAI API, and no other external cloud or service is involved. OpenAI also maintains stringent security standards, comparable to Salesforce, and does not train its models on user prompts or ground data.
Despite these considerations, there are significant security advantages to enabling AI. With AI, administrators can detect security vulnerabilities, quickly document compliance issues, discover technical debt, and enhance the quality of organizational assets. For comprehensive details about Metazoa’s Zero Trust architecture and OpenAI API usage, please see our Security Policy document at:
https://www.metazoa.com/privacy-and-security/
AI Preferences
Snapshot has a dedicated panel for AI preferences, providing flexible options to address varying corporate policies. In the Snapshot application, go to Snapshot Menu > Snapshot Preferences and navigate to the third tab to configure OpenAI usage. You can choose from three different modes:
- Don’t Use OpenAI Services
- Use Metazoa Free OpenAI Key
- Use Your Corporate OpenAI Key
Selecting Use Your Corporate OpenAI Key is generally recommended, as it allows you to monitor usage and costs while preventing unexpected service interruptions. Average expenses for most customers are expected to be under $100 per month, though large-scale org transformations with extensive AI usage might incur higher costs. For more information on obtaining an OpenAI key:
https://platform.openai.com/api-keys
The Metazoa Free OpenAI Key is also available to all customers, although its usage is limited for processes that might result in massive resource consumption. The primary limitations are:
- Calculating SOQL Filters: Unlimited Use
- Explaining Migration Errors: Unlimited Use
- Explaining Deployment Errors: Unlimited Use
- Explaining Metadata Assets: Unlimited Use
- Metadata Studio Prompt Engineering: 50 Uses / Individual
- Metadata Studio Prompt Automation: 5 Uses / Individual
- Intelligent Search Assistant: 100 Uses / Individual
- Description Enrichment: 100 Uses / Individual
Additionally, there is a checkbox that allows you to Mask Personal Information. When enabled, any sensitive fields—like email addresses or phone numbers—are scrambled before sending them to the API, then unscrambled upon receiving responses. This feature helps preserve data privacy during AI-assisted tasks.
Finally, users can select from different OpenAI Models. While the default model (gpt-4o-mini) offers a cost-effective option with ample context window, advanced models like gpt-4o and others may provide improved response quality. However, each model comes with its own token and character limit constraints. It’s essential to choose a model that best aligns with your project needs.
SOQL Filters
Snapshot often requires administrators to select or manipulate data records, especially during data migration or when preparing datasets. While there are multiple ways to carry out record selection, one of the most efficient—and also one of the most syntax-heavy—is SOQL filters. AI assistance significantly reduces the learning curve and time spent on code syntax with the new “Create Filter” button, which uses AI to convert a textual description into a correct SOQL string.
This feature includes an interactive testing interface that lets you validate your query results before finalizing. By leveraging AI, even teams with minimal SOQL experience can easily set complex filters, ensuring only the desired records are returned.
Explaining Errors
Deployment errors from the Salesforce Metadata API can be cryptic and time-consuming to decipher. To expedite troubleshooting, Snapshot’s AI functionality offers an Explain Error Messages option. Right-click on an error in the deployment error list and receive an explanation, along with recommended solutions for resolving the issue.
Similarly, for Data Migration, if you encounter an unexpected error during the dataset migration process, you can right-click the error message to open the same Explanation interface. This capability helps resolve issues promptly, saving time for both administrators and developers.
Explaining Assets
Many administrative tasks involve viewing or editing metadata assets such as XML-based Custom Objects or text-based Apex Classes. Snapshot now enables you to right-click any asset and select Explain Asset, which displays a specialized interface offering prompts that delve into different aspects of the asset.
Administrators can choose default prompts or create custom prompts to meet specific organizational needs. Common prompts include:
- Automated Documentation
- Best Practices
- Compliance Review
- Functional Specification
- Technical Debt
These prompts provide targeted insights—whether you need a quick documentation summary, highlight compliance gaps, or reveal potential vulnerabilities. By incorporating this feature into your regular development workflow, you can maintain cleaner and more understandable metadata assets.
Metadata Studio
Metadata Studio serves as a robust AI-driven prompt engineering platform for Salesforce administrators and developers who want to automate or standardize repeated org management tasks. It comes pre-loaded with 60 default prompts, fully customizable to match any internal guidelines or policies. The interface consists of three tabs:
- Prompt Selection:
Quickly choose a default prompt from the list and click Submit to run it. This provides an immediate and convenient way to explore various operations, like scanning an Apex Class for technical debt or reviewing Compliance aspects of a Custom Object.
- Prompt Engineering:
In this tab, you can fine-tune existing prompts or develop new ones using “messages.” Each prompt begins with a “System” message and can be followed by any combination of “User” or “Assistant” messages. Content for each message can be metadata assets, data records, XML sections, or plain text. You can even mark some content as Required, enforcing that end users select specific assets before running a prompt. This tab also includes options like:
- Dynamic Compression: Automatically compresses assets for large metadata or data records, ensuring the prompt stays within model token limits.
- Dynamic Documentation: Generates thorough descriptions of metadata elements for a richer AI response.
- Assets Limits: Restricts selection to certain asset types—like Apex test classes—promoting targeted usage and compliance with organizational standards.
- Prompt Automation:
Once you have developed or selected a suitable prompt, you can apply it to multiple metadata assets across your Salesforce org. For instance, you may create a prompt to document all Apex Classes and combine the details into a unified report. Alternatively, you might automate the generation of node descriptions for multiple Flows consecutively. After completion, these changes can be compiled into a single report, multiple reports, or even saved as a local Developer Project. If the prompt involves data records, the results can be saved for Dataset Migration, or if the prompt involves metadata assets, immediate Metadata Deployment is possible.
Metadata Studio effectively streamlines repeated tasks, whether they relate to security reviews, compliance checks, or large-scale documentation updates.
Intelligent Search
Intelligent Search expands your admin team with an AI-based “thinking partner” that can analyze and interpret your Salesforce org’s data and configuration. There are four key categories of information Intelligent Search can reference:
- Data Records: Select an Account and any connected Opportunities or Contacts, then ask about next steps to close a deal or highlight potential data issues.
- Object Schema: Provide an object’s schema information, and the AI can propose validation rules, new fields, picklist values, or formula fields.
- Metadata Types: Highlight multiple assets such as Validation Rules, Flows, or Apex Classes, then let the AI search for conflicts, best practices, or security flaws.
- Org Management: Ask broader administrative questions that require referencing Metazoa’s whitepapers, best practices, and documentation. Intelligent Search will provide recommendations for managing your org’s complexity.
By focusing on these foundational data sets, Intelligent Search offers unparalleled clarity about your org’s structure and health. This feature is particularly valuable during security assessments or mass updates.
Description Enrichment
The Description Enrichment report allows administrators to systematically review and update the description fields of Salesforce metadata. Over 160 metadata types contain description fields, which are often neglected or updated sporadically. This oversight can hamper future documentation, especially in complex or highly regulated environments.
With the Description Enrichment tool, you can see which metadata assets lack descriptions or contain inadequate ones. You can then review and revise these descriptions directly within Snapshot. Additionally, AI can be used to generate new descriptions automatically. Different prompts can be specified for different metadata types, ensuring context-specific descriptions. Once generated, you can quickly deploy these changes using Snapshot’s deployment tools.
This uniform process for description management helps maintain both compliance and clarity, fostering an environment where your entire team understands the purpose and function of each metadata asset.
Conclusion
With AI-driven tools like Intelligent Search and Metadata Studio, Metazoa strives to transform how Salesforce orgs are managed. By merging robust administrative features with an AI-backed zero-trust security framework, Snapshot delivers new levels of efficiency and optimization. From automatically documenting metadata assets to detecting hidden security vulnerabilities, these AI capabilities bolster every aspect of an administrator’s workflow.
Moving forward, these AI features will continue to evolve as technology advances, offering an expanding range of capabilities for admins, developers, and architects alike. Snapshots’s unique combination of Zero Trust architecture, data privacy measures, and comprehensive AI usage controls ensures that organizations can adopt these features confidently and responsibly.
Congratulations! 