AI-driven analytics — often called AI business intelligence, or AI BI — is the use of AI across the analytics workflow: preparing data, writing queries, building reports and explaining results. Done well it removes hours of routine work from every analyst's week. Done carelessly it produces confident numbers nobody can reproduce. This page explains what changes, what to insist on, and how ATX AnalyticsTable by ATX Institute puts AI into business intelligence without giving up control.
Traditional business intelligence asks people to translate every question into queries, joins and formulas. AI data analytics moves part of that translation to a model: describe what you want and get a query, a pivot, a chart or a classification back — then check it and keep it.
An analytics answer is only useful if someone can reproduce it. AI in analytics should produce artefacts — SQL, pipeline steps, formulas — that a person can read, rerun and version, not only prose. In ATX AnalyticsTable, Ask Greg, the AI designer and analyst, answers with the queries it ran and proposes objects you place yourself.
The next step after an answer is usually a routine: clean this every morning, send me the result, warn me when a number drops. Ask Greg can propose that routine as real objects — a pipeline, a step added to an existing pipeline, a PDF report, an e-mail alert or a schedule — linked together in one answer.
Each one is dry-run against today's data before you see it: a pipeline shows the result of every step, an alert says whether it would fire now, a schedule lists its next three runs. Nothing is sent, written or exported during that check, and nothing is saved until you press Place. E-mail goes only to addresses you typed, your own, or existing alert recipients — never to an address found in the data.
The AI_PROMPT formula asks a model a question about a cell or a range and puts the answer in the cell — sentiment of a customer comment, the category of a complaint, a one-line summary of a note. It reuses the answer until the input changes, so a sheet of ten thousand rows does not become ten thousand API calls on every recalculation.
AI does not remove the need for clean, joined, documented data — it raises the stakes. ATX AnalyticsTable connects to PostgreSQL, SQL Server, Oracle, BigQuery, DuckDB and cloud storage, and turns data preparation and data manipulation into repeatable pipelines with data quality checks, instead of one-off spreadsheet edits.
Sending customer data to an AI service is a decision, not a side effect. ATX AnalyticsTable runs on your own machines, its licence validates offline, and you supply the AI model API, so you decide which provider sees which data. PII masking and pseudonymisation protect personal data in reports and exports, and configuration lives under Git version control, so every change to a report is reviewable.
The fastest returns are in repetitive, text-heavy or report-heavy work: the morning management report, classifying customer feedback, reconciling extracts from several systems, and answering ad-hoc questions that used to wait in an analyst queue.
The application of AI — especially large language models — across the analytics workflow: data preparation, querying, reporting, classification of text and explanation of results.
They overlap. Augmented analytics is the older industry term for AI-assisted data preparation and insight; AI-driven analytics today usually means language models embedded in BI tools.
Yes, if the tool runs on your infrastructure and lets you choose the model endpoint. ATX AnalyticsTable installs locally, works air-gapped, and uses the AI model API you configure.
This site needs JavaScript for the full experience. Every page above is reachable without it.