
Industries
Salesforce Service Cloud orgs, service-ops and RevOps teams, B2B support organizations
Business Need
Teams drowning in scattered signals and list views with no single, explainable answer to "what should I do next?", Directive turns that into one prioritized, auditable queue
Requires
Directive_User, Directive_Manager, Directive_Admin, Directive_AI, and Directive_Sensitive_Actions permission sets; 6 custom objects and 5 Custom Metadata Types deployed together (Apex tests read seeded CMDT records); a UI Bundle-capable org on API v67.0
Compatible With
Agentforce (contextual explain/summarize/classify/recommend), Salesforce CLI/DX tooling, Custom Metadata-driven configuration, React + TypeScript UI Bundles
Support Level
Standard
Reviews Rating
0

Full operational loop: Detect, Understand, Prioritize, Explain, Execute, Measure, built on 6 custom objects and 5 Custom Metadata Types DirectivePriorityService: a fully explainable, decomposable scoring engine, every point traceable via the PriorityExplanation UI Typed action registry (Directive_Action_Definition__mdt) with custom-permission checks and idempotent audit rows (Directive_Action_Run__c) before any DML React + TypeScript multi-framework UI Bundle with 8 routes: Today, Work, Work Detail, Customers, Approvals, AI Activity, Insights, Settings Agentforce employee agent scoped to explain, summarize, classify, and recommend, never to execute unrestricted DML Mock-data dev mode (VITE_DATA_MODE) for exploring the full UI before wiring live Salesforce data via a repository/adapter seam
Service-ops teams triaging which case, account, or signal to act on next without digging through list views Managers reviewing prioritized, explainable work queues instead of raw CRM records Admins configuring rules, personas, actions, and scoring via Custom Metadata without code changes Agentforce-assisted summarization and recommendation on top of an auditable action trail Teams evaluating the UI against realistic mock data before connecting live Salesforce data
Requires
Directive_User, Directive_Manager, Directive_Admin, Directive_AI, and Directive_Sensitive_Actions permission sets; 6 custom objects and 5 Custom Metadata Types deployed together (Apex tests read seeded CMDT records); a UI Bundle-capable org on API v67.0
Compatible With
Agentforce (contextual explain/summarize/classify/recommend), Salesforce CLI/DX tooling, Custom Metadata-driven configuration, React + TypeScript UI Bundles
API Version
Salesforce API v67.0 required (UI Bundle will not deploy below it). Lightning only; no Classic or Experience Cloud support.
Integrations Required
Salesforce CLI (sf); Node.js 22+; Custom Metadata Types deployed alongside Apex classes (tests read seeded records); Directive_User/Manager/Admin/AI/Sensitive permission sets; optionally @salesforce/platform-sdk when switching the UI from mock data to live Salesforce data
Setup Time
Roughly 30-60 minutes for a working baseline: authorize the CLI against your org, run the deploy script to build the UI bundle and deploy all metadata, assign the Directive_User permission set, then open the Directive Lightning app.
Version
1.0.0
Listed On
Latest Release
Supported Features
Signal detection and trigger-based work-item creation; explainable priority scoring via DirectivePriorityService; typed action registry with custom-permission gating; idempotent audit rows for every action run; React/TypeScript UI with 8 routes covering the full operational loop; mock-data dev mode with a live-data adapter seam
Languages
TypeScript, Apex, CSS, Shell, HTML, JavaScript
Package Type
Metadata (Unmanaged)
Edition Support
Requires API version 67.0 for the UI Bundle to deploy; no namespace. No Data Cloud or Big Objects required.
Mobile Compatible
No, the UI Bundle app is a desktop Lightning experience; no mobile-specific layout is declared
Every action passes through a typed action registry, custom-permission checks (Directive_Sensitive_Actions / Directive_AI_Actions), and writes an idempotent audit row (Directive_Action_Run__c) before any DML. Agentforce is scoped to explain, summarize, classify, and recommend, it never executes unrestricted DML. Deploy scripts deliberately assign only Directive_User by default so negative permission tests keep passing.
Data Cloud Required
No
SCREENSHOTS
See it in action
A closer look at the interface, dashboards, and day-to-day workflows.



