Glossary

The four-stage loop that describes how Appice processes every customer signal. Sense ingests the signal; Decide determines the next best action; Act fires the action; Learn records the outcome and feeds it back. Every interaction runs through this loop, typically in under 100ms.

Any data point that indicates a meaningful change in a customer's context, intent, or state — a transaction, a login, an app event, a usage pattern shift. Signals are the inputs that start the SDAL loop.

The unified customer profile assembled from behavioural, transactional, and demographic data across all systems. Every decision is made against this enriched record, not raw event data.

The process of matching device IDs, email addresses, phone numbers, and account numbers to a single persistent customer profile — the prerequisite for accurate cross-channel decisioning.

A continuous, ordered sequence of timestamped events flowing from source systems into Appice in real time. Every customer behaviour generates an event that can trigger an immediate decision.

A customer attribute derived by calculation — average transaction value, days since last login, total spend in 30 days. Computed traits are updated continuously as new signals arrive and are available to the decisioning engine instantly.

The single optimal action to take for a specific customer at a specific moment — whether that is a retention offer, a product recommendation, a service alert, or no action at all. The decisioning engine selects the NBA by evaluating rules, propensity scores, and business objectives simultaneously.

An if-then condition that evaluates a signal against customer attributes, segment membership, or propensity scores. Rules enforce eligibility, suppression, frequency caps, and compliance guardrails before any action is fired.

An AI-generated probability that a customer will take a specific action — convert, churn, lapse, upgrade. Propensity scores are applied at decision time to rank eligible actions and select the most likely to succeed.

A dynamic group of customers who share defined characteristics. In Appice, segments update in real time as customer signals change — a customer enters a "high churn risk" segment the moment their behaviour crosses the threshold, not at the next batch refresh.

A per-customer limit on how many times an action can fire within a defined window — per day, per week, per channel. Enforced automatically at decision time to prevent over-communication.

The elapsed time between receiving a customer event and returning a decision. Appice targets sub-100ms decision latency — fast enough to inject a personalised offer into an active mobile session before the user navigates away.

A condition that blocks an action from firing for a specific customer — for example, excluding customers who are in a complaint state, already hold the product being offered, or have opted out of a channel.

The delivery of the decided action across the right channel — push notification, SMS, email, WhatsApp, in-app message, call centre prompt — based on the customer's channel preference and availability at the moment of decision.

A defined event or condition that initiates an action. In Appice, triggers replace scheduled sends — the action fires the moment the customer signal arrives, not on a calendar date.

An action delivered while the customer is in an active session — in the mobile app, online banking, or on a call. In-session delivery has the highest conversion rate because the customer is present and attentive.

Automatic routing to a secondary channel when the primary channel is unavailable or undeliverable — for example, falling back from push notification to SMS if the customer's app is inactive.

An immutable record of every decision — the signal that triggered it, the customer context at that moment, the rules that fired, the action selected, and the outcome. Required for regulatory traceability in banking, insurance, and healthcare.

The process of assigning credit for a conversion or outcome to the specific decision and action that caused it. Appice resolves attribution per customer in real time — delivery, engagement, and conversion are linked in a single journey log.

A controlled experiment that runs two different decision rule sets, offers, or next best actions simultaneously — with outcome measurement built in. Used to continuously improve decision logic without rebuilding campaigns.

The process by which the result of each action — converted, ignored, bounced — is recorded back into the system and used to update propensity models and decision rules. This is what makes the SDAL loop self-improving over time.

The agentic orchestration layer that coordinates decisions across the SDAL loop — evaluating context, compliance constraints, and offer eligibility simultaneously, under 100ms, without human intervention.

The Central Notification Service for backend-triggered campaigns. Sending systems pass opaque customer tokens; Appice handles identity resolution and channel delivery. Preserves PII outside the calling system.

AI systems that autonomously perceive context, make decisions, and execute actions without requiring step-by-step human instruction. In Appice, agentic AI orchestrates multi-step workflows across the full SDAL loop in a single transaction.

The requirement that data be stored and processed within a specific geographic jurisdiction. Appice supports on-premise and private cloud deployment for Tier 1 banks and regulated organisations with in-country residency requirements.

Every automated decision in Appice is logged with a human-readable reason code — which rules fired, which score applied, which conditions were met. Required for regulatory compliance in banking, insurance, and healthcare.