AI and data glossary
Plain definitions of the terms used across this site, written by the team that builds the systems.
- AI agent
A language-model system that carries out a task in steps, calling tools and business systems such as ERP or CRM, rather than only answering a question.
Related: Generative AI
- AI readiness assessment
A review of whether an organization can put AI to productive use. Snowfox reviews five dimensions together: data, technology, strategy, people and culture, and governance and processes.
Related: AI Readiness Assessment
- Big data
Data whose volume or speed exceeds what a single database handles comfortably, so it needs distributed processing such as Apache Spark, NoSQL stores or cloud warehouses, often in real time.
Related: Big data consulting
- Causal machine learning
Models built to estimate what an action will cause, such as the effect of a price change on demand, rather than only what tends to happen alongside it.
Related: Price optimization for a retailer
- Computer vision
Models that interpret images and video: recognition, object detection, segmentation and video analysis, applied to inspection, document processing and security.
Related: Computer vision
- Cross-price elasticity and cannibalization
How a price change on one product moves the sales of others. Cannibalization is the costly case: a promotion that grows one product by taking sales from another in the same portfolio.
Related: priceGURU
- Customer and offer segmentation
Grouping customers by behavior so each group gets the offer most relevant to it. Machine learning finds groups that conventional segmentation misses.
Related: segMENTOR
- Data engineering
Building and maintaining the pipelines, warehouses and analysis layers that make data available and trustworthy for reports and models.
Related: Data engineering
- Data governance
The rules for who owns data, who may use it, how long it is kept and how it is protected: security, retention and privacy you can defend.
Related: Data-driven strategy
- Data pipeline and ETL
An automated sequence that moves data from source systems to where it is used. ETL (extract, transform, load) is the classic form: pull the data, clean and reshape it, load it into a warehouse.
Related: Data engineering
- Data strategy
A statement of which decisions data should improve and what architecture, governance and tooling are needed to make that possible.
Related: Data-driven strategy
- Data warehouse and data mart
A data warehouse is a central store of cleaned, organized data built for analysis. A data mart is a smaller slice of it shaped for one subject or team.
Related: Data engineering
- Entity and relationship extraction
Finding the named things in a text (people, companies, dates, amounts) and how they relate, for example who grants which powers to whom in a contract. Relationships are the harder part.
Related: Automated PDF extraction using IDP
- Exit-intent
A message or offer shown at the moment a visitor is about to leave a website, used to recover the visit.
Related: segMENTOR
- Generative AI
Models that produce text, code, images or audio. In business systems it usually means large language models used for assistants, agents, document work and code migration.
Related: Generative AI
- Intelligent document processing (IDP)
A pipeline that turns documents such as PDF contracts into structured data: reading the page, understanding its layout, extracting entities and relationships, and validating the result.
Related: Automated PDF extraction using IDP
- Large language model (LLM)
A model trained on very large amounts of text that can read and write natural language and code. It is the engine behind most generative AI systems.
Related: Generative AI
- Machine learning (ML)
Software that learns patterns from data instead of following hand-written rules, and uses them to predict, classify or recommend.
Related: AI consulting
- MLaaS
Machine learning as a service: using pre-built models that a provider hosts and maintains, paid as an operating cost, instead of building and owning the models.
Related: MLaaS
- MLOps
Machine learning operations: the practices for deploying, monitoring and improving machine-learning models in production with a repeatable process. It is to models what DevOps is to software.
Related: MLOps
- Model drift
The gradual loss of accuracy of a model in production as customer behavior, prices or products move away from the data it was trained on. Monitoring and retraining are the remedy.
Related: MLOps
- Natural language processing (NLP)
Techniques that let software read, classify and produce text and speech: classification, entity extraction, summarization, generation and translation.
Related: Natural language processing
- Price elasticity
How much the quantity sold of a product changes when its price changes. An elasticity of -2 means a 1% price increase reduces units sold by about 2%.
Related: priceGURU
- Reinforcement learning
A way of training a system by trial and feedback: it tries options, observes the result and shifts toward what works. Used to keep improving which offer or creative is shown.
- Remote sensing
Measuring land and crops from a distance with satellite or drone imagery, used in agriculture to map cultivated areas, monitor crop health and forecast yields.
Related: Agritech
- Retrieval-augmented generation (RAG)
A design in which the system first retrieves relevant passages from your own documents and then has a language model answer from them. It is how enterprise search and internal assistants stay grounded in company data.
Related: Generative AI
Start with the problem, not the technology
Tell us what is slow, expensive or invisible in your operation. If AI or better data is the right answer, we will show you where. If it is not, we will tell you that too.
Not ready to scope a project?
Start with the AI Readiness Assessment. Five dimensions, reviewed together, and an honest read on where you actually are.
Take the assessment