AI Glossary A–Z — Enterprise AI, Agentic AI & CX Terms

160+ terms · Curated by our practice leads · Updated as the market evolves

Showing 1–30 of 168 terms

A17 terms on this page

After-Call Work (ACW): Post-interaction tasks (wrap codes, notes, dispositions). Auto-summary and auto-disposition are the highest-ROI early CX AI wins.
Agent Assist: Real-time AI suggestions surfaced to a human agent during a customer interaction — answers, next-best-action, sentiment cues, after-call summary.
Agentic AI: AI systems that pursue goals autonomously by planning multi-step actions, calling tools and APIs, and revising their approach based on feedback — rather than only responding to a single prompt.
AI Agent: A software entity built around an LLM that perceives context, reasons over a task, invokes tools, and produces an outcome — often on behalf of a user or another agent.
AI Engineer: A software engineer who ships production LLM systems — prompt design, RAG, tool integration, evals, cost/latency tuning. Distinct from ML engineer.
AI Gateway: A managed proxy that routes model calls, enforces quotas, redacts PII, logs prompts, and applies policy across multiple LLM providers.
AI Product Manager: Owns discovery, scoping and lifecycle of AI features — writing evals as acceptance criteria, managing model/prompt versions, and quantifying business impact.
AI Solutions Architect: Designs end-to-end AI systems: models, retrieval, tools, integration, security, cost. The senior engineering role that bridges enterprise architecture and applied AI.
AIOps: Applying machine learning to IT operations data (logs, metrics, traces) to detect anomalies, correlate incidents and automate remediation.
Alignment: The discipline of making an AI system's behavior match human intent, values and organizational policy — both at training time and at runtime.
Anomaly Detection: Identifying data points or events that deviate from expected patterns — fraud, outages, churn signals, quality defects.
API Gateway (LLM): A managed proxy that routes model calls, enforces quotas, redacts PII, logs prompts, and applies policy across multiple LLM providers.
Artificial General Intelligence (AGI): A hypothetical AI capable of performing any intellectual task a human can, across domains. Not a shipping product category today; treat vendor claims with scrutiny.
Artificial Intelligence (AI): Software systems that perform tasks normally requiring human intelligence — perception, reasoning, language, decision-making — by learning patterns from data.
AUC (Area Under Curve): A single-number summary of a classifier's ability to separate classes across all thresholds; useful for imbalanced problems like fraud or churn.
Autonomy Level: A taxonomy (advisor → co-pilot → supervised agent → autonomous agent) describing how much human oversight a workflow requires.
Average Handle Time (AHT): Mean duration of a customer interaction including hold and wrap. A primary CX KPI reshaped by agent-assist and automation.

B7 terms on this page

Batch Inference: Running many model calls asynchronously at lower cost — used for backfills, document processing and offline analytics.
Benchmark: A standardized task suite (e.g., MMLU, GSM8K, HELM) used to compare model quality on reasoning, coding, safety or domain knowledge.
Bias (Model Bias): Systematic error in model output that disadvantages a group or skews decisions — introduced through data, labeling, objectives or deployment context.
BPO (Business Process Outsourcing): Delegating end-to-end business processes (contact center, claims, finance) to a third-party operator; increasingly restructured around AI-augmented agents.
Bring Your Own Model (BYOM): A platform pattern that lets enterprises deploy fine-tuned or proprietary models inside a vendor's runtime instead of using only the vendor's default model.
Build vs. Buy vs. Fine-tune: The core AI portfolio decision: adopt a vendor product, build on foundation-model APIs, or invest in fine-tuned/proprietary models. Rarely a single answer per enterprise.
Business Case (AI Business Case): A structured argument tying an AI initiative to measurable value, cost, risk and time-to-impact — the artifact that clears funding gates.

C6 terms on this page

Caching (Prompt/Response): Storing repeat prompts or key/value tensors to cut latency and cost; supported natively by most frontier providers.
CCaaS (Contact Center as a Service): Cloud-native contact center platform — routing, IVR, WFM, analytics, digital channels — delivered as a subscription (e.g., Amazon Connect, Genesys Cloud, NICE CXone, Five9, Google CCAI).
Chain-of-Thought (CoT): A prompting or training pattern where a model produces intermediate reasoning steps before its final answer, improving accuracy on math, planning and multi-hop tasks.
Chief AI Officer (CAIO): Executive accountable for enterprise-wide AI strategy, portfolio, governance and value realization. Increasingly a peer to CIO, CDO and CISO.
Chunking: Splitting source documents into passages sized for retrieval and the model's context window. Chunk size and overlap are top-3 RAG quality levers.
Citation / Attribution: Linking each generated claim back to the source passage that supports it — non-negotiable for regulated and knowledge-worker use cases.

All 168 terms, A–Z

Every entry has its own page with a plain-English definition, why it matters to enterprise buyers, and where it fits in a delivery program.

Go deeper

Take the terms into practice.

Each glossary entry maps to a guide, benchmark or working session in our library.

Guides

The AI Glossary for enterprise leaders — A to Z.

All 168 terms, A–Z

Take the terms into practice.

Pillar guides & reference architectures

Benchmark data on adoption, containment and ROI

Translate the vocabulary into a program