The business glossary for finance, held in the semantic layer that computes your statements.
Keboola’s business glossary is a governed metric catalog for finance teams: ~197 definitions — formula, source, owner — held in the same semantic layer that computes your P&L and balance sheet. Your dashboards and Kai, Keboola’s AI assistant, read the same definitions. EBITDA means one thing in every report and every AI answer.
Trusted by 1,000+ companies
The Problem
Group EBITDA becomes the sum of four different concepts.
Same metric, three definitions.
EBITDA means one thing in your SAP, another in the FP&A model, and a third in the board deck. Every reconciliation meeting starts by agreeing on which version is correct.
New hires inherit broken logic.
Definitions live in a controller’s head or a stale Confluence page. When that person leaves, the logic leaves with them. Every new analyst reverse-engineers the formula from report output — and gets it slightly wrong.
AI gives confident wrong answers.
When AI queries an undefined metric, it interpolates. In our discovery calls, one multi-entity CFO showed us an AI-generated year-end cash position that was about 30% off. Two analysts ask the same question, get different numbers. Both technically justified. Neither defensible.
Auditors cannot follow the logic.
A metric that cannot be defined unambiguously cannot be audited reliably. The auditor asks for the definition. The answer is a spreadsheet, a long explanation, and a 48-hour wait.
Business Glossary
Defined once. Enforced everywhere.
The semantic layer is the guardrail — every number Kai returns is computed from your definitions, not interpreted from them. When Kai is asked for EBITDA, it executes your definition, not an interpretation of it.
0
Definition per metric — enforced across all systems
0%
AI-query consistency — Kai and your BI read the same definitions
Zero
“What does this mean?” meetings about your own numbers
How it works
Surface
Connect your data catalog, ERP metadata, and FP&A models. The first step is an inventory — we surface every metric, KPI, and account grouping in active use, including ones nobody knew existed.
Define
Finance owns each definition — formula, source, owner, approved use. Each one lives in the semantic layer, not in a document. Versioned and change-logged from day one.
Enforce
Definitions are wired into the model itself — the semantic-layer P&L and balance sheet compute from them. Every dashboard, report, and AI query — from Kai or any agent connected over MCP, Copilot, Gemini, or ChatGPT — resolves against the same definition. The definition is the calculation, not a comment on it.
Monitor
Every definition change lands in a live audit and governance feed — who changed what, when, and why, reconstructable by your auditor. Drift is caught in review, not in the board meeting.
Companies that fixed it
Companies that ended the definition debate.
Every executive had their own point person for a number.
75% of decisions are now data-driven, up from effectively zero. Keboola unified 20+ source systems — including broken and outdated ones — into one governed layer, ending the point-person-per-number dependency.
Unified financial definitions across 9 regulatory environments.
Once definitions were unified, 25% of all HQ reporting could be automated — the team shifted from data pushing to enabling intelligence.
Why Keboola
Most BI tools let you define metrics inside dashboards. That definition lives in the dashboard, not in the data.
When someone builds a second dashboard, they redefine it. Keboola governs the definition at the data layer — every tool that reads from it, including every agent connected via MCP, computes the same governed answer. And each definition is portable SQL your team owns. No black box. No vendor lock-in on your business logic.
Questions & answers
Frequently Asked Questions
Everything finance, IT, and procurement will want to know — up front.
Resources for CFOs
Insights & conversations
for finance leaders.
Whitepapers, ebooks, implementation playbooks — and conversations with finance leaders operating in multi-entity, PE-backed, and post-acquisition environments.
