Glossary

Names in bold are the names you see in the app.

Your data

Term	Meaning
Conversation	The unit of data Dimension Labs reads: one chat, ticket, call, review or survey response, with every message in it. All the messages of one conversation share a sessionId.
Conversation data	Text a person wrote or said: chats, calls, tickets, reviews, survey answers. It goes under Unstructured data and is enriched into Dimensions.
Business data	Rows and columns such as revenue, orders or accounts. It goes under Structured data, is queryable within about a minute of upload, and is never enriched. It comes in by file upload only today; an API for it is on the roadmap.
Data Sources	The page in the app where your data comes in. It has two panels: Unstructured data and Structured data.
Unstructured data	The panel in Data Sources for conversation data: your integrations and your conversation file uploads.
Structured data	The panel in Data Sources for business data tables.
Source	One item in Data Sources: an integration or an upload under Unstructured data, or a table under Structured data.
Live integration	A source that sends conversations as they happen: a one-click integration, or a universal API job sending to /track. It is enriched automatically every day.
Metadata	The extra columns that come with your conversations, such as rating, store, plan or channel. Stored in structured_metadata.
Session	The grouping set by sessionId: every message with the same sessionId belongs to one conversation.
sessionId	Required. The id that groups messages into one conversation.
userId	Required. The end user’s id, on their messages and on the bot’s or agent’s replies to them. If you don’t have one, copy the sessionId into it. Never an email address, phone number or name.
incoming	Required. true for a message from the customer or end user, false for a message from your agent, bot or company.
text	Required. The words of one message.
dimensionlabs_timestamp	Required. When the message was sent, as a 13-digit Unix time in milliseconds. Messages in one session need to be at least 1 second (1,000 ms) apart.
conversationId	Recommended. Keeps one conversation together across a handoff, for example from a bot to a human agent.
intent	Optional, and only for an intent-based chatbot. An object: { "name": "TRACK_ORDER", "confidence": 0.94 }.
platformJson	An optional nested object of extra fields sent with a message. Never put personal information in it.
partition_key	The platform’s key for one conversation, used to join tables. Spelled partition_key in Dimension tables and PARTITION_KEY in structured_metadata.
transcripts	The table holding every message, one row per message.
structured_metadata	The table holding the metadata you sent with your conversations, one row per conversation.

Dimensions and enrichment

Term	Meaning
Dimension	One typed field extracted from each conversation: a category, a yes or no, a score, or a verbatim quote.
Prompt	The configuration that defines a set of Dimensions, with its instructions. Each prompt writes its own table.
Standard Dimensions	The general Dimensions every workspace has available, such as sentiment and contact reason. Called System Prompts in the app. They run every day only when Dimension Labs turns on daily enrichment for your workspace (ask support); otherwise, run a backfill.
System Prompts	The app’s name for the standard Dimensions. Their results are in the system_session_prompt table.
system_session_prompt	The table that holds the standard Dimensions, one row per enriched conversation.
Custom Dimensions	Dimensions you design for your business, in your own prompts.
Verbatim	A Dimension that quotes the customer’s exact words behind a label, so the label can be checked.
none_detected	The value a Dimension returns when the model read the conversation and found nothing for that field. Different from an empty value, which means nothing was returned.
Enrichment	Running prompts over conversations to write their Dimensions.
Backfill	A run that applies a prompt to conversations you already have. Needed for a file upload and for anything sent to /trackhistorical. Start it from the Backfill tab. It typically shows up within the hour, depending on how much data is being enriched.
Backfill tab	The tab in the app where you start a backfill.
Prompts tab	The tab in the app that lists every prompt in your workspace.
Prompt Designer	A setting in the agent. Turn it on in the agent’s settings to design and save prompts with the agent.
Meaning layer	The whole structure the agent works on: your original conversations, the Dimensions extracted from each one, and your business data, in one schema you can query together.

The agent and outputs

Term	Meaning
Agent	The analytics agent you talk to in plain English. It reads your tables, runs queries and statistics, and builds outputs.
Chat	One conversation with the agent. Your past chats are listed in the left rail.
Project	A workspace for a piece of work: its files, its standing instructions (AGENTS.md) and its chats.
AGENTS.md	A file at the top of a project that the agent reads as a standing instruction in every chat of that project.
Analysis plan	A written research plan the agent drafts with you: the question, the data, the steps and the shape of the output. Saved as a file so it can be run again on new data.
File Manager	The agent’s view of your project’s files. It has the button that saves a file as an artifact.
Final Reports	The folder in a project where the agent saves finished deliverables.
Artifacts	Published reports, dashboards and decks. Publishing gives a file a link anyone can open, and you can revoke the link. Publish from the File Manager or the Artifacts page.

APIs and admin

Term	Meaning
Universal API	The API that receives conversation messages from any system, at tracker.dimensionlabs.io. Also called the tracker. See Universal API.
/track	The universal API endpoint for new messages, sent as they happen. Enriched automatically every day.
/trackhistorical	The universal API endpoint for older messages. Use it for anything older than an hour, then run a backfill.
Export API	The API that returns your enrichments as a zip file, by day. Filter by integration ID and date range. Its key is in Settings. See Export API.
Settings	Where you find your Export API keys.
User management	The page where you invite users and see who has accepted.