Transitioning from a traditional, UI-heavy quoting system to a headless, agentic revenue engine requires a shared vocabulary. This glossary defines the core pillars of the methodology to ensure seamless alignment across all project workstreams.
| Term | Definition & Framework Context |
|---|---|
| Agent Action | A specific capability registered within the Agentforce brain. It connects the LLM's understanding of intent to an execution mechanism, such as an OmniStudio Integration Procedure or an Invocable Apex class. |
| Commercial Archetype | The structural blueprint or "shape" of a transaction (e.g., Subscription-Linked, Usage-Based, One-Time). It defines the strict JSON data payload parameters required by downstream systems. |
| Commercial Matrix | The complete repository of deterministic business rules, pricing waterfalls, regional legal constraints, and margin ceilings that govern the AI's boundaries and prevent execution errors. |
| Discrete Intent | The precise objective isolated by the LLM from natural language input (e.g., a customer asking to "add 50 seats mid-term") that directly triggers an automated execution pathway. |
| Headless Payload | The structured JSON data model generated completely by the AI agent from natural language text, containing all transactional variables necessary to execute a backend operation without human screen clicks. |
| Intent Domain | The macro industry vertical or ecosystem (e.g., SaaS, Manufacturing, Telecom) that establishes the overarching conversational vocabulary and fundamental revenue rules the AI must understand. |
| Intent Cluster | A grouped category of business or user objectives within a domain (such as "Subscription Modifications") that maps to a collection of related Commercial Archetypes. |
| Matrix Discovery | The agile investigative methodology used by architects to interview business subject matter experts (SMEs). Its sole purpose is to isolate and document the hard rules and guardrails needed to feed the Commercial Matrix. |
| Orchestration Routing Logic | The technical rules determining whether an AI agent can execute an action silently via backend APIs or if it must pause the payload and route it through a human-in-the-loop approval workflow. |
| Platform-Guided Design (Pre-Gap Steering) | The consulting practice of using discovery sessions to immediately steer business desires into standard out-of-the-box platform engines (like Salesforce RLM's Product Selling Models) instead of defaulting to custom development. |
| Traceability Ledger | The 12-phase compliance framework that audits a transaction at every stage—from initial human chat prompt, through AI extraction, down to the final financial subledger entry—ensuring absolute data integrity. |