The language of agentic retail

An AI CRM Manager is an autonomous AI worker that owns a retailer's customer-lifecycle workflows end-to-end: reasoning, remembering, deciding, and executing 1:1 for every customer, on top of the CRM, CDP, and marketing-automation stack the retailer already runs.

An AI Decision Engine is the reasoning layer that sits above a retailer's data and execution stack and outputs, per customer and continuously, the next best commercial move, with its rationale attached.

AI Decisioning is the use of AI to determine what should happen next for a customer in a lifecycle workflow. Traditional AI decisioning platforms decide for a micro-segment; 1:1 AI Decisioning decides for a single customer.

The AI Decision Layer is the missing layer between a retailer's data (CDP, warehouse, event stream) and their execution surfaces (CEP, MAP, ecommerce, app). It reasons per customer, decides what should happen next, and outputs an execution-ready instruction.

Theory of Mind for Retail is Replenit's methodology for modeling each customer's intent, beliefs, and likely next move, then deciding the commercial action that fits, the way an attentive human merchant would.

A Golden Decision Event is Replenit's execution-ready output unit. It packages a 1:1 customer decision (what should happen, for which customer, why now, with which product, in which tone, toward which commercial outcome) as a portable, inspectable JSON structure ready to fire into any CRM, CDP, or marketing-automation platform.

In Replenit, Skills are category-specific commercial capabilities Maestro applies to a decision, the way a seasoned category merchant would. A Skincare Regimen Planner Skill lets Maestro reason like a skincare expert; a Cosmetic Routine Builder Skill lets it build routines like a beauty advisor.

Golden Memory is Replenit's persistent, enriched understanding of every customer, product, and brand, updated continuously as new events arrive. It is the layer that lets Maestro maintain continuity between decisions instead of treating each event as isolated.

In retail, the Concierge Experience is a 1:1 customer-facing communication that reads like advice from an assigned personal shopper: what was picked, why, what was deliberately left out, and the value each product brings. It is the antonym of a "personalized" campaign or a "you might also like" recommendation.

In retail, a Lifecycle Workflow is a sequence of 1:1 decisions and actions that pursues a customer outcome: replenishment, cross-sell, engagement, winback, product churn, promotion, or product substitute. Maestro owns each workflow end-to-end per customer.

Agentic CRM is the emerging category of CRM systems that operate with autonomous AI agents rather than human-configured workflows. The category splits into two levels: task agents that help a marketer do their work faster, and worker agents, like an AI CRM Manager, that own an outcome themselves.

An AI Workforce for Retail is a stable of autonomous AI workers, each a hire-able role that owns a retail outcome end-to-end. The first role is an AI CRM Manager for the customer lifecycle; future roles will own other functions.