Glossary
The vocabulary, without the jargon.
Short definitions of the terms this site uses. Where a term relates to a Doo product, it links there.
A
- AI agent
- A language model that can use tools — look something up, call an API, run a workflow — and use the result to decide its next step. See AIDoo
- AI gateway
- A single, governed entry point through which applications reach language models, so access, cost and policy are managed in one place. See AIDoo
- AI governance
- The inventory, policies, approvals and audit applied to AI systems: what exists, who owns it, what it may do and what it has done. See GovernanceDoo
- Audit trail
- A chronological record of who did what to which asset. A chained audit trail makes any later alteration evident.
C
- Calculated column
- A column whose values are computed from other columns by a formula. See FlowDoo
- Connector
- The component that lets a platform read from or write to a particular system or data source. See FlowDoo
- Control plane
- The shared layer that manages identity, policy, metadata and audit for everything that runs on a platform. See GovernanceDoo
- Cross-filtering
- Selecting a value in one chart to filter the other charts on the same dashboard. See VizzDoo
D
- Data catalog
- An inventory of an organization's datasets with their meaning, owner and classification. See GovernanceDoo
- Data governance
- The practice of making data owned, understood, protected and fit for use through agreed roles, policies and controls. See GovernanceDoo
- Data integration
- Bringing data from separate systems together in a consistent form so it can be used as one. See FlowDoo
- Data lineage
- The path a piece of data took: where it came from, what changed it and where it went.
- Drill-down
- Moving from a summarised figure to the more detailed data behind it. See VizzDoo
E
- ELT
- Extract, load, transform: loading raw data into a destination first and transforming it there.
- ETL
- Extract, transform, load: reading data from a source, reshaping it, and writing the finished result to a destination. See FlowDoo
- Explainable AI
- Methods and practices that help people understand why a model produced a result. It supports trust and review; it does not by itself make a model correct. See MLDoo
F
- Feature
- An input variable a machine learning model uses to make a prediction. See MLDoo
G
- Guardrails
- Rules applied to what goes into and comes out of an AI system to keep it within policy. See AIDoo
M
- Model governance
- Controlling how machine learning models are documented, reviewed, approved, deployed and monitored. See GovernanceDoo
- Model monitoring
- Watching a deployed model's inputs and outputs over time to notice when its behaviour changes. See MLDoo
- Model Passport
- MLDoo's record for a model: owner, training data, performance, features, explainability, limitations, risk rating, approval and deployment status. See MLDoo
- Model registry
- A versioned store of models and the evidence about each one. See MLDoo
R
- RAG
- Retrieval-augmented generation: giving a language model relevant passages from your own content so that it answers from them and can cite them. See AIDoo
- RBAC
- Role-based access control: granting permissions to roles, and roles to people, rather than permissions to individuals.
S
- Semantic layer
- A shared set of business definitions — measures and dimensions — that every report uses, so a metric means one thing. See VizzDoo
- SHAP
- A widely used method for estimating how much each feature contributed to an individual prediction. See MLDoo
T
- Tool calling
- A language model requesting that a defined function be run with specific arguments, and using the result. See AIDoo
W
- Workflow automation
- Running the steps of a process automatically, in order, with the checks and hand-offs the process needs. See FlowDoo
One platform. A better way to build with data and AI.
Start with one use case. Expand across the ecosystem.