Business intelligence (BI) is the work of turning operational data into figures people can decide on — connecting the sources, preparing the data, defining the measures, and delivering them as dashboards, reports and alerts. Most BI tools assume that work happens in someone else's cloud. This guide walks through each layer of a BI stack, what to ask of it in a regulated business, and how ATX AnalyticsTable by ATX Institute covers it on machines you control.
A BI tool is judged less by its charts than by the chain behind them. If any link is manual — a copy-paste, an export someone forgets — the dashboard at the end is only as reliable as that step. A complete BI stack covers:
ATX AnalyticsTable connects to local DuckDB and SQLite files, PostgreSQL, Microsoft SQL Server, Oracle and Google BigQuery, and imports Parquet, CSV, JSON and Excel from Azure Blob Storage, Google Cloud Storage and S3-compatible object stores. Credentials sit in an encrypted vault behind the operating system's keychain, and transport security is explicit — PostgreSQL from DISABLE through VERIFY-FULL, SQL Server certificates validated by default.
Results can be cached into the project's own DuckDB database on demand, so paging through a cloud warehouse result is billed once rather than once per page.
Every reliable BI number starts as a query. ATX AnalyticsTable puts a SQL editor and a visual query builder beside the grid, keeps a saved-query library the whole project can point at, and lets a parameter such as a date range change every report at once instead of editing SQL each morning. Query results land in the sheet as pinned queries — live previews for exploring, static pins for production — and _SQL formulas push SUM, AVG and COUNT down to the database instead of pulling millions of rows into the sheet.
Self-service BI means the people who own a question can answer it without filing a ticket. The risk is that every analyst builds a private copy of the truth. ATX AnalyticsTable keeps self-service inside guard rails: a semantic layer gives raw tables friendly business names and relationships, and every query, pipeline, schedule and export definition is a plain-text file in the project folder — so a change is a reviewable Git diff, not a surprise in next month's figures.
Dashboards are their own tabs: a canvas of KPI tiles, filters, pivots, charts and titles, with each widget bound directly to a connection rather than to a copied range. Cascading filters let one selection drive the whole view, scheduled refresh keeps the numbers current before the meeting, and a dashboard design exports as JSON without any row data.
Between the source and the report sits preparation: sorting, de-duplicating, mapping values, cleaning text, masking personal data and running SQL. In ATX AnalyticsTable that is a pipeline — an ordered, versioned recipe re-applied on every refresh, with an append-only run log — that can also write its result into DuckDB, PostgreSQL or BigQuery, overwriting or appending.
Most people consume BI as a report that arrives, not a dashboard they open. Numbered schedules run strictly in order — refresh, then recalculate, then export — and deliver dated Excel files, branded PDF reports or e-mails with the table attached. Alert rules send only when a condition is met, with a cooldown and an optional recovery message.
Ask Greg, the AI analyst in ATX AnalyticsTable, profiles columns, runs SELECT-only SQL against your own database and proposes saved queries, pivots, charts, KPIs, dashboards and reports. Its SQL is held to a stricter contract than a person's — classified rather than pattern-matched — and nothing it proposes is placed until you accept it.
A short checklist that separates BI software you can defend in an audit from software that only demos well:
Business intelligence (BI) is the set of tools and practices that turn raw business data into trusted measures, dashboards and reports that people use to make decisions.
BI focuses on reliable, repeatable reporting of what is happening — dashboards, KPIs and scheduled reports. Data analytics is broader and includes exploratory and predictive work. ATX AnalyticsTable covers both in one project: ad-hoc SQL and AI analysis next to the scheduled reports built from them.
Yes. ATX AnalyticsTable is installed on your own machines, connects to your databases directly, validates its licence offline, and produces dashboards, PDFs, exports and e-mails locally — nothing is sent to a reporting cloud.
It covers the same core BI jobs — connections, SQL, dashboards, scheduled reports — but runs on your own infrastructure with a spreadsheet grid business users already know, and keeps every definition as a file under Git. Compare the plans on the pricing page; dashboards are included in every plan except Corporate PRO.
This site needs JavaScript for the full experience. Every page above is reachable without it.