AI Glossary
Plain-language definitions of the AI, automation, and Arabic NLP terms that come up when building AI systems for Gulf businesses.
core aiAI Agent
An AI agent is an AI system that pursues a goal by taking a sequence of actions — calling tools, looking up information, and deciding what to do next — rather than only producing a single text reply.
core aiAI Readiness
AI readiness is the degree to which a business's data quality, existing processes, tools, and team skills are prepared to successfully adopt and benefit from AI, rather than stalling on a pilot that never reaches production.
arabic nlpArabizi
Arabizi is a informal way of writing spoken Arabic using Latin letters and numbers (such as '3' for ع and '7' for ح), widely used by GCC and Egyptian users in texting, WhatsApp, and social media instead of Arabic script.
voice aiASR (Automatic Speech Recognition)
ASR (Automatic Speech Recognition) is AI technology that converts spoken audio into written text, forming the listening half of any voice agent or transcription system.
core aiContext Window
A context window is the maximum amount of text — measured in tokens — that an AI model can read and hold in working memory at one time, including the prompt, conversation history, and any retrieved documents.
arabic nlpDialect ASR
Dialect ASR is automatic speech recognition trained and evaluated specifically on regional spoken Arabic dialects, such as Egyptian or Gulf Arabic, rather than only Modern Standard Arabic.
accessibilityEAA (European Accessibility Act)
The EAA (European Accessibility Act) is an EU law, enforceable since June 28, 2025, that requires many digital products and services — including e-commerce sites, banking apps, and e-books — sold in the EU to meet accessibility standards for people with disabilities.
core aiEmbeddings
Embeddings are numerical representations of text (or images and audio) that place similar meanings close together in a mathematical space, letting AI systems compare content by meaning rather than exact keyword matches.
core aiFine-Tuning
Fine-tuning is the process of taking an existing pretrained AI model and continuing its training on a smaller, focused set of your own examples, so it reliably adopts a specific tone, format, dialect, or task without being trained from scratch.
core aiGuardrails
Guardrails are the rules, filters, and checks placed around an AI model — on both what goes in and what comes out — that block unsafe, off-topic, off-brand, or factually ungrounded responses before they ever reach a customer.
core aiHuman-in-the-Loop (HITL)
Human-in-the-loop is an AI system design where a person reviews, approves, or can override the AI's output at a defined checkpoint before it takes effect, rather than letting the system act fully on its own.
core aiLLM Evals (Large Language Model Evaluations)
LLM evals are structured, repeatable tests that score an AI model or agent's outputs against a set of criteria — accuracy, tone, safety, dialect handling — using a fixed set of test questions, so quality can be measured objectively instead of judged by spot-checking a few chats.
core aiModel Monitoring
Model monitoring is the ongoing tracking of a live AI system's real-world performance — accuracy, response time, cost per conversation, escalation rate, and failure patterns — after it is deployed, so problems are caught from the data rather than from a customer complaint.
arabic nlpMSA vs. Dialect
MSA (Modern Standard Arabic, or Fusha) is the formal, standardized Arabic used in writing, news, and official documents, while dialect (Amiya) — such as Gulf, Egyptian, or Levantine Arabic — is the everyday spoken language that varies significantly by country and region.
core aiPrompt Engineering
Prompt engineering is the practice of designing and refining the instructions given to an AI model so its output is accurate, consistently formatted, and aligned with a specific business need.
core aiRAG (Retrieval-Augmented Generation)
RAG (Retrieval-Augmented Generation) is a technique where an AI system first retrieves relevant passages from your own documents or database, then uses them as grounding context to generate its answer — instead of relying only on what the underlying model memorized during training.
voice aiTTS (Text-to-Speech)
TTS (Text-to-Speech) is AI technology that converts written text into natural-sounding spoken audio, forming the voice half of any speaking voice agent.
core aiVector Database
A vector database is a specialized database that stores content as numeric embeddings and retrieves results by semantic similarity — meaning — instead of exact keyword matches.
accessibilityWCAG (Web Content Accessibility Guidelines)
WCAG (Web Content Accessibility Guidelines) is the internationally recognized standard, published by the W3C, that defines how to make websites and apps usable by people with visual, auditory, motor, or cognitive disabilities.
whatsapp aiWhatsApp Business API
The WhatsApp Business API is Meta's platform for medium and large businesses to send, receive, and automate WhatsApp messages at scale through approved technology providers, rather than the manual WhatsApp Business app.