Glossary
Terms used in this manual, alphabetised.
| Term | Definition |
|---|---|
| Action | A tool the agent can call during a conversation — display a form, render a card, show a booking calendar. Auto-instantiated by skill configuration; there's no standalone Actions UI in the current product. |
| Activity | A first-class event on a prospect record: chat session, meeting booking, contact form submission, profile change, integration sync. Visible on the Prospects screen's History tab. |
| Allowance | How much of something your plan includes each cycle — conversations, messages and document pages. Shown on the Billing & Usage page. When conversations or messages run out, the agent stops until the allowance resets or you upgrade. |
| Agent | The AI chatbot. Configured with a profile (profession, personality, response length, goal), a greeting, a set of skills, and the conditions and nudges on them. One team can have multiple agents. |
| Agent type | A template that determines an agent's default skill set and prompts. The only generally-available agent type today is sAIlsbot; the others (ChatBot, FAQBot, Lighthouse, Manifest, SupportBot) are marked Coming Soon. |
| Attribute | A structured property extracted from product-shaped content (e.g. waterproof_rating, plan_tier). Used by the Product Finder skill. Not relevant for services-led tenants. |
| Book Meeting skill | The skill that shows a booking calendar in the chat, drawing on your connected Cal.com account's availability. Requires a connected Cal.com account and an event type chosen for the skill — both set up from Integrations. |
| Cal.com | The booking provider the platform integrates with for Book Meeting. Connected once per team over OAuth; the skill then books against one of the account’s event types. |
| Channel | A deployment target type. Currently Chat widget is the only available channel. |
| Chip | One of the tappable buttons under a teaser. Tapping it sends its wording as the visitor's first message and hands the conversation to a skill, so the wording is fixed rather than editable. A teaser offers a pool; you pick 1–3. |
| Chunk | A self-contained passage of ingested content with its own vector embedding, used for retrieval. The chunker is AI-driven and tries to preserve semantic coherence. Product pages chunk into a fixed set of categories, one chunk each, with an overview always produced. |
| Consent prompt | The widget's Do you want us to remember you next time? prompt, answered Yes or No. Asks whether an identifier may be kept in the visitor's browser, so they are recognised on a return visit. Appears after the visitor submits details (booking, contact form) or after four messages. It does not gate processing: what they submitted, and what the agent works out, reach you either way. |
| Contact Request skill | The skill that displays an inline contact form and submits captured details to the prospect record. |
| Content | Ingested knowledge — scanned web pages or uploaded documents. Each piece of content moves through parse → classify → chunk → attribute extract → embed before being usable by the agent. |
| Conversation Quality | A 1–5 score from the summariser estimating how well the agent handled a session. Averaged across sessions on the Analytics screen. |
| Custom property | A property you defined yourself (rather than one of the platform's seeded built-ins). Lives in the Custom tab of the Properties screen. |
| Deployment | A configured instance of an agent on a channel. Has its own theme, content copy, allowed domains, and embed snippet. One agent can have multiple deployments. |
| Deployment key | A unique UUID that identifies a deployment. Used in the widget loader URL. |
| Embed snippet | The single-line <script> tag you paste into a website to install the widget. Bound to a specific deployment via the deployment key. |
| Firecrawl | The third-party service that crawls a site to find its pages, and fetches them when you import. |
| Condition | A rule standing in front of a skill, deciding who may enter it. While the agent can't yet tell, it asks one qualifying question and the answer decides. A condition may also name where everyone else is sent; that destination stays open to everyone. Applies however the visitor arrives, including a direct request. Configured on the skill it governs. |
| Conversation inputs | The per-skill list of things a skill should learn from the visitor, each set to a level — mandatory (pursued until answered), optional (asked when it would help), or inferred (never asked, worked out from signals). |
| Greeting | The first message the agent sends, plus the suggested-question chips and chat-input hint that appear with it. Configurable per agent and overridable per deployment. |
| Informational skill | The built-in skill that answers questions about the company, services, policies, etc. using the knowledge base. Always on and not listed on the Skills screen — there is nothing to configure. FAQ content gets a lift in the ranking when a general question is searched. |
| Insights | The right-hand panel in the playground that fills in after a session ends — summary, sentiment, outcome, resolution, quality, content analysis. Mirrors what the summariser computes for every session. |
| Integration | A connected third-party service — HubSpot, Zoho CRM, or Cal.com. Authenticated via OAuth 2.0; managed from the team’s Integrations screen. |
| Knowledge base | All the content ingested into a team — the corpus the agent retrieves from during conversations. |
| Lead Score | A 0–100 score the platform computes from a prospect's property values, segment matches, and activities. Higher = warmer. |
| Manage data | The visitor-facing panel for reviewing and erasing personal data, reachable from the widget's menu icon. The platform-side equivalent of a GDPR data-subject access request. |
| MQL | Marketing Qualified Lead. A built-in segment name; the criteria that define it are tenant-configurable. |
| Nudge | A rule that leans the agent toward a target skill when a prospect matches a segment or a one-off test. A preference, not a rule the agent must obey. Evaluated deterministically, with a per-session lifecycle (pending → fired → resolved). Configured on the skill it leans toward. |
| Test | The one-off check a condition or nudge carries instead of a segment — one property, one comparison, one value, written on the card rather than saved under a name. A card takes exactly one. There is no ordering between rules: a nudge lifts a skill's ranking and nothing forces the agent to take the top one. |
| Nudge engine | The deterministic, non-LLM component that evaluates conditions and nudges against the session's property values every turn and decides how the reply ends — a nudge toward a skill, a qualifying question, or just the answer. |
| Organisation | The top-level tenant. Everything in the platform belongs to an organisation. |
| Topic | What a page is about — product, audiences, company or knowledge. Shown as the Topic column on the Content list, and used to steer how the agent searches your content. Separate from Page type, which is what the page is for. |
| Page type | What a page is for, as distinct from what it is about — a product page, a listing, an FAQ, pricing, contact, a guide, a case study, a page for a particular audience, an about page, your home page, or your legal and checkout pages. Worked out from the page's own content and markup, never from its URL. Exactly one teaser per page type; legal and checkout pages carry none. |
| Plan | What your organisation is subscribed to — Free, Launch, Growth or Scale, or a trial. Decides your allowances and which features are available. Managed by an owner on the Billing & Usage page. |
| Playground | The in-dashboard test environment for trying conversations against your agent. Same chat pipeline as the production widget. |
| Profile (agent) | The agent's persona configuration — profession, personality, response length, goal — under agent settings. For visitor classification, see Segment. |
| Prompt materialisation | The server-side process of filling prompt templates with team / agent / deployment values, run automatically when configuration changes. |
| Property | The schema for a kind of value the agent can capture (e.g. Company Name, Current Challenge). Configured per team on the Properties screen. Either built-in or custom. |
| Property value | A specific value the agent captured about a prospect (e.g. Company Name = Breezee). |
| Prospect | A visitor who has interacted with the chat widget. Each prospect carries the captured property values, the conversation history, and any contact information they shared. |
| Reciprocal Rank Fusion (RRF) | The scoring technique the retrieval engine uses to combine vector similarity, title match, and content text match into a single ranked result list. |
| Remember me | The visitor's Yes answer to the remember-me prompt. It keeps a marker in their browser, so they are recognised when they return and later visits join the same prospect record. It gates nothing else: contact details, properties, the lead score and the session summary are kept whichever way they answer. |
| Resolution Rate | The percentage of sessions the platform classified as Resolved — i.e. ones where the visitor reached a meaningful outcome. |
| Retrieval policy | One of default, product_specific, or informational. Determines how the search engine weights vector / title / content scoring and whether the FAQ fast-path applies. |
| Segment | A named pattern of property criteria that classifies a prospect — MQL, Hot Lead, Tire Kicker, etc. Segments are consumed by conditions and nudges, and shown as chips on the Prospects screen. Deleting one deletes the rules that use it. |
| Session | A single chat conversation. Created when a visitor opens the widget or starts a playground session; closed by End Session or a long idle period. |
| Session actions menu | The three-dot menu in the playground that holds Manage data and End Session. |
| Skill | A modular AI capability you attach to an agent. Configurable out of the box: Advisory Guidance, Contact Request, Deliver Download, Product Finder, Book Meeting, Order Form, Web Store Purchase, Emergency Contact Request — plus built-ins that need no setup (Greeting, Informational, End of Chat, and fallbacks for off-topic or unclear messages). |
| Slot extractor | The LLM call that pulls structured values (property values) out of the visitor's message on every turn. Runs in parallel with intent and tool planning. |
| Summariser | The background LLM call that runs after each chat turn and at session end. Updates the running session summary and (at end) produces the full Insights output. |
| Teaser | The small bubble that appears beside the launcher to invite a visitor into a conversation, before they click anything. Chosen per page type on the agent's Engagement tab; whether teasers fire at all, and how soon, is set per deployment under Proactive engagement. |
| Team | A workspace inside an organisation. Holds agents, content, prospects, and deployments. The URL paths call this concept projects; the UI calls it team. |
| Tool planner | The LLM call that decides per-turn whether to search the knowledge base and which retrieval policy to use. |
| Trial | The 14 days a new organisation starts with: every Growth feature on a smaller allowance. If no plan is chosen by the end, the organisation moves to the Free plan and the agent stops taking conversations; nothing is deleted. |
| Unstructured.io | The third-party document-parsing service used to extract text from uploaded PDFs, Word docs, and similar non-Markdown formats. |
| Page view | One load of a page carrying the widget — the denominator for engagement analytics. Not a count of people: one person reading three of your pages counts three times. Shown on the Analytics screen as Page Views (previously Website Visitors). |
| Widget | The embeddable chat interface that visitors interact with on your website. Renders as an iframe inside a launcher button. |