AI Demo Glossary: 40+ Terms Every Sales and Product Team Should Know

Most glossary pages are SEO plays. This one actually came from our internal onboarding doc — we kept explaining the same terms to new hires and decided to publish the whole thing. If a term is defined here, it's because someone at RaykoLabs or one of our customers needed it explained at least twice.

AI-powered demo automation sits at the intersection of artificial intelligence, browser automation, voice technology, and sales strategy. That intersection produces a lot of jargon. Some comes from the AI research community, some from sales operations, and some is emerging alongside the category itself.

---

AI Demo Agent

An autonomous AI system that delivers interactive, personalized product demonstrations without human involvement. An AI demo agent combines product knowledge, browser automation, and conversational AI to navigate the actual product, explain features, answer questions, and adapt the experience to each prospect in real time. Unlike static tours or recorded videos, an AI demo agent takes actions and makes decisions during the demo. For a deep dive, see our complete guide to AI demo agents.

ASR (Automatic Speech Recognition)

The technology that converts spoken language into text. ASR systems use deep learning models trained on large datasets of human speech to transcribe audio with high accuracy. In voice-enabled demos, ASR is the first step in processing a prospect's spoken question or request. Also referred to as speech-to-text or STT. RaykoLabs uses Deepgram for ASR — chosen for its streaming capability and accuracy across accents and background noise conditions.

Browser Automation

The programmatic control of a web browser — clicking elements, typing text, navigating between pages, scrolling, and interacting with web applications without manual input. In AI demo agents, browser automation enables the agent to navigate the actual product interface in real time. RaykoLabs uses Playwright for browser automation and Browserbase for cloud-hosted browser sessions, which eliminates infrastructure headaches. For more detail, see browser automation for live AI demos.

Buyer Intent

Signals that indicate a prospect's likelihood to purchase. In AI demos, buyer intent is derived from what a prospect asks about, which features they explore, how long they engage, and what objections they raise during a demo session. Voice-enabled demos generate richer intent data than any other format because prospects articulate their needs and concerns in natural language — no more guessing from click patterns.

Click-Through Demo

A guided, interactive product experience built from screenshots, HTML captures, or sandbox environments where the prospect advances by clicking highlighted elements. Click-through demos follow a predetermined path and cannot adapt to open-ended questions. Platforms like Navattic, Storylane, and Walnut specialize in this format. They work for structured overviews but lack the flexibility of AI-driven demonstrations.

Context Detection

The ability of an AI demo agent to understand the current state of the product interface during a demo. Context detection involves reading the DOM, identifying visible elements, understanding what data is displayed, and using this information to make intelligent decisions about narration and navigation. At RaykoLabs, context detection is the first of three navigation layers — followed by navigation planning and LLM integration — that let the agent reliably find its way through any product interface.

Conversational AI

A category of artificial intelligence focused on enabling natural, human-like dialogue between humans and machines. Conversational AI combines natural language processing, machine learning, and dialogue management to understand user input and generate appropriate responses. In AI demo agents, conversational AI powers the ability to have a real-time, adaptive conversation with prospects during a product demonstration.

CRM Integration

The connection between an AI demo agent and a customer relationship management system such as Salesforce, HubSpot, or similar platforms. CRM integration allows demo session data — including conversation transcripts, features explored, questions asked, and engagement metrics — to flow directly into lead records, enabling sales teams to prioritize and personalize follow-up.

Demo Automation

The use of technology to deliver product demonstrations without requiring a live human presenter. Demo automation encompasses a spectrum of approaches, from simple recorded videos to sophisticated AI demo agents. The goal is to scale the demo experience beyond the capacity constraints of human sales teams.

Demo Environment

A dedicated instance of a product configured specifically for demonstrations. A demo environment is populated with realistic sample data, configured to showcase key workflows, and maintained separately from production systems. AI demo agents navigate the demo environment during sessions, so its stability and data quality directly impact the demo experience.

Demo Fatigue

The declining effectiveness of product demonstrations caused by prospects being overwhelmed with too many demos during an evaluation cycle, or by repetitive, generic demo experiences that fail to address specific needs. AI demo agents combat demo fatigue by delivering personalized, on-demand experiences that respect the prospect's time. For strategies to address this, see how to combat demo fatigue.

Demo No-Show

An event where a prospect schedules a product demonstration but fails to attend. Demo no-show rates in B2B sales typically range from 20 to 40 percent, representing lost opportunity and wasted sales engineering time. On-demand AI demos eliminate the no-show problem entirely. For more on why prospects disappear, see why prospects ghost your demos.

DOM (Document Object Model)

The programming interface for web documents that represents the structure of a page as a tree of objects. AI demo agents interact with the DOM to understand what is displayed on screen, identify clickable elements, and execute navigation actions. DOM awareness is a critical capability that enables the agent to operate on live product interfaces rather than scripted simulations.

DOM Awareness

The ability of an AI demo agent to read, interpret, and interact with the Document Object Model of a web application in real time. DOM awareness allows the agent to identify buttons, links, form fields, and displayed data, which is essential for both navigation and for generating accurate verbal descriptions of what is on screen.

Embedding

In the context of AI, an embedding is a numerical representation of text, audio, or other data in a high-dimensional vector space. Embeddings allow AI systems to measure the semantic similarity between pieces of content. In AI demo agents, embeddings are used to match prospect questions to relevant product knowledge and to enable retrieval-augmented generation.

Generative AI

A category of artificial intelligence that creates new content — text, images, audio, code, or video — based on patterns learned from training data. In AI demo agents, generative AI powers the creation of natural-sounding responses to prospect questions, ensuring each answer is contextually appropriate rather than pulled from a static script.

Guided Tour

A structured, step-by-step walkthrough of a product that leads the user through a predetermined sequence of screens and features. Guided tours are typically non-interactive beyond basic navigation clicks and do not adapt to user questions or interests. They serve as a lightweight introduction but lack the depth and personalization of AI-driven demos.

Hallucination

In AI, a hallucination occurs when a language model generates information that is factually incorrect, fabricated, or unsupported by its training data or knowledge base. In AI demo agents, preventing hallucination is critical — the agent must never invent features, misstate capabilities, or make unsupported claims about the product. Robust implementations use grounding techniques and guardrails to minimize this risk.

Interactive Demo

A product demonstration format that allows the viewer to interact with the product rather than passively watching. Interactive demos range from simple click-through tours to fully conversational AI demo agents. The level of interactivity directly correlates with engagement, information retention, and conversion rates.

Knowledge Base

The structured repository of information that an AI demo agent uses to answer questions and narrate the product experience. A knowledge base typically includes product documentation, feature descriptions, pricing information, competitive positioning, objection handling, and use case libraries. The quality and completeness of the knowledge base directly determines the quality of the demo experience.

Latency

The time delay between a user action and the system's response. In voice-enabled demos, latency refers to the total round-trip time from when a prospect finishes speaking to when they hear the agent's response and see the product begin to move. Low latency — under two seconds — is essential for natural conversational flow. At RaykoLabs, we target under 800ms to first audio. Anything above that and prospects start talking over the agent, which creates a cascade of problems.

Lead Scoring

The process of assigning a numerical value to each lead based on their likelihood to become a customer. AI demo agents enhance lead scoring by providing behavioral data from demo sessions — what features the prospect explored, what questions they asked, how long they engaged, and what buying signals they exhibited — alongside traditional firmographic and demographic data.

LLM (Large Language Model)

A deep learning model trained on vast amounts of text data that can understand and generate human-like language. LLMs are the reasoning engine behind AI demo agents, responsible for interpreting prospect questions, generating responses, determining navigation actions, and maintaining conversational context. Examples include models from the GPT, Claude, and Gemini families.

Live Demo

A product demonstration conducted in real time by a human sales representative or sales engineer, typically via a video conferencing platform. Live demos offer the highest degree of personalization and human connection but are constrained by scheduling availability and headcount. AI demo agents complement live demos by handling initial evaluation and qualifying interest before a live conversation.

Navigation Planning

The process by which an AI demo agent determines the sequence of actions required to navigate from the current state of the product interface to a desired destination. When a prospect asks to see a specific feature, the agent must plan a path through menus, pages, and clicks to get there. At RaykoLabs, navigation planning is the second of three layers — sandwiched between context detection and LLM integration — in our three-layer navigation system.

NLP (Natural Language Processing)

A branch of artificial intelligence focused on enabling computers to understand, interpret, and generate human language. NLP encompasses tasks like intent recognition, entity extraction, sentiment analysis, and language generation. In AI demo agents, NLP is used to understand what prospects are asking and to generate appropriate, contextually relevant responses.

On-Demand Demo

A product demonstration available to prospects at any time, without requiring scheduling or human involvement. On-demand demos address the fundamental scalability challenge of traditional scheduled demos. AI demo agents are the most sophisticated form of on-demand demo, offering the interactivity and personalization of a live experience with the availability of self-serve content.

Personalization

The adaptation of a demo experience based on the prospect's role, industry, company size, stated interests, or behavior during the session. In AI demo agents, personalization occurs dynamically — the agent adjusts its narrative, feature emphasis, and navigation path based on signals detected during the conversation. This is a different model from the static segmentation used in click-through demos.

Playwright

An open-source browser automation framework developed by Microsoft that enables programmatic control of Chromium, Firefox, and WebKit browsers. RaykoLabs uses Playwright as its browser automation layer — it navigates product interfaces, clicks elements, fills forms, and captures screen state. We chose Playwright over Puppeteer for its cross-browser support and more reliable handling of single-page applications with dynamic content loading.

Product-Led Growth (PLG)

A business strategy where the product itself drives customer acquisition, conversion, and expansion. In PLG organizations, self-serve experiences like free trials and interactive demos are primary growth levers. AI demo agents align with PLG strategies by providing a guided, intelligent product experience that converts website visitors into active users without requiring sales involvement.

Prompt Engineering

The practice of designing and optimizing the instructions given to a large language model to achieve desired outputs. In AI demo agents, prompt engineering determines how the agent introduces features, handles objections, maintains conversational tone, and decides when to suggest scheduling a call with a sales rep. Effective prompt engineering is critical for demo quality and consistency.

Prospect

A potential customer who has expressed interest in a product, typically by visiting a website, downloading content, or engaging with a demo. In the context of AI demo agents, the prospect is the person interacting with the agent — asking questions, exploring features, and evaluating whether the product meets their needs.

RAG (Retrieval-Augmented Generation)

A technique that enhances a large language model's responses by retrieving relevant information from an external knowledge base before generating a response. In AI demo agents, RAG ensures that the agent's answers are grounded in accurate, up-to-date product information rather than relying solely on the LLM's training data. This reduces hallucination and improves response accuracy.

Real-Time Demo

A product demonstration that occurs live, with actions and responses generated instantaneously as the prospect interacts with the system. AI demo agents deliver real-time demos where the product navigation, voice responses, and adaptive behavior all happen during the session rather than being pre-recorded or pre-scripted.

Sales Engineering

A specialized role within B2B sales organizations responsible for conducting product demonstrations, answering technical questions, and supporting the sales process with product expertise. AI demo agents augment sales engineers by handling initial product evaluations at scale, freeing human SEs to focus on high-value, complex deal support.

Sandbox Demo

A product demonstration conducted in an isolated environment where the prospect can interact with the actual product without affecting real data or systems. Sandboxes provide authentic product experiences but typically lack guidance. AI demo agents operating within sandbox environments combine the authenticity of a sandbox with the guidance and narrative of a structured demo.

Screen Capture Demo

A demonstration format created by recording or capturing screens of the product and assembling them into a guided sequence. Screen capture demos are static representations that cannot show live data, respond to input, or adapt to questions. They are simpler to create than live demos but offer a far less compelling experience than AI-driven alternatives.

Self-Serve Demo

A product demonstration that prospects can access and complete independently, without requiring assistance from a sales representative. Self-serve demos range from embedded videos and click-through tours to AI demo agents. The trend toward self-serve reflects buyers' preference for evaluating products on their own terms and timeline.

Session Recording

The capture of a complete demo session — including screen activity, voice conversation, navigation paths, and interaction data — for later review and analysis. RaykoLabs uses rrweb for session recording, which captures DOM changes at a granular level rather than recording video. This produces smaller files, pixel-perfect playback, and the ability to search within sessions. Sales teams use these recordings to understand prospect needs and prepare for follow-up conversations.

Speech-to-Text (STT)

See ASR. Speech-to-text is the process of converting spoken audio into written text. In voice-enabled demos, STT is the input channel that allows prospects to communicate with the AI demo agent using natural speech rather than typing. Streaming STT implementations process audio in real time, enabling more natural conversational dynamics.

Streaming

The continuous transmission of data in small chunks rather than waiting for the complete payload before delivery. In AI demo agents, streaming is used throughout the pipeline — streaming STT processes audio as it arrives, streaming LLM inference begins generating responses before the full input is processed, and streaming TTS begins playback before the complete audio is synthesized. Streaming is the primary technique for minimizing conversational latency.

Text-to-Speech (TTS)

The technology that converts written text into natural-sounding spoken audio. TTS is the output channel that gives AI demo agents a voice. RaykoLabs uses Cartesia for TTS — it produces speech with human-like intonation, pacing, and expressiveness while supporting streaming output. In voice-enabled demos, TTS converts the LLM's generated response into audio that is streamed to the prospect's browser. The choice of TTS engine directly impacts how "human" the demo feels.

Token

In the context of large language models, a token is the basic unit of text that the model processes. A token can be a word, part of a word, or a punctuation mark. LLMs have token limits for both input (context window) and output (response length), which affect how much information the agent can consider and how detailed its responses can be. Token usage also directly impacts computational cost.

Turn-Taking

The conversational mechanism that determines when one party stops speaking and another begins. In voice-enabled demos, effective turn-taking requires the agent to detect when the prospect has finished speaking — distinguishing between a natural pause and the end of a statement. Poor turn-taking results in the agent interrupting the prospect or waiting too long to respond, both of which degrade the experience.

Vector Database

A database optimized for storing and querying high-dimensional vector embeddings. In AI demo agents, vector databases power the retrieval component of RAG by enabling fast similarity searches across the product knowledge base. When a prospect asks a question, the system converts it to a vector and finds the most semantically similar knowledge entries to include in the LLM's context.

Voice Agent

An AI system that communicates with users through spoken language, using speech-to-text for input and text-to-speech for output. In the context of demos, a voice agent is an AI demo agent that supports voice-based interaction, allowing prospects to speak naturally rather than type. Voice agents create a more engaging and accessible demo experience compared to text-only alternatives.

Voice Demo

A product demonstration where the primary interaction modality is voice. The prospect speaks to the AI demo agent, which responds with spoken audio while simultaneously navigating the product. Voice demos combine the convenience of a phone conversation with the visual richness of a screen-shared product walkthrough.

WebSocket

A communication protocol that provides full-duplex, persistent connections between a client (the prospect's browser) and a server. In AI demo agents, WebSockets enable real-time, bidirectional streaming of audio data, navigation commands, and state updates. This persistent connection is essential for the low-latency, real-time interaction that voice-enabled demos require.

---

Why this terminology matters

These terms aren't academic. As AI demo agents become a standard component of B2B sales tech stacks, the ability to evaluate solutions and measure outcomes depends on shared vocabulary.

Sales leaders evaluating AI demo platforms need to ask informed questions about latency, hallucination prevention, and DOM awareness. Product marketers need to understand the difference between a click-through demo and a voice-enabled AI demo agent. Revenue operations teams need to understand CRM integration and lead scoring capabilities when building their stack.

Here's a hot take: most teams buying demo automation tools can't define half the terms on this page. That's not a criticism — the category is new and the terminology is borrowed from multiple fields. But the teams that take time to understand the underlying technology make better buying decisions. We've seen it firsthand.

For the complete picture on how these technologies work together, read our guide to AI demo agents, our explanation of how the RaykoLabs demo agent works, or our complete guide to AI demo automation.

This glossary is a living document. As the category evolves and new capabilities emerge, new terms will follow.