Frequently Asked
Questions

Three deployment models:

• Multi-tenant SaaS: Keboola manages all infrastructure. Available stacks: AWS US/EU, Azure EU, GCP US/EU
• Multi-tenant with BYODB: Data stays in customer-owned Snowflake or BigQuery while Keboola manages orchestration
• Single-tenant: Complete Keboola stack within customer cloud (AWS, Azure, GCP). Maximum control for regulated industries.

Supported backends:

• Snowflake: Default backend with full SQL support, dynamic warehouse sizing, native workspaces, BYODB
• BigQuery: Full SQL support, native datatypes, direct workspace access, BYODB
• Additional via connectors: Redshift, Synapse, Oracle, Exasol, Teradata, Firebolt

Learn more at help.keboola.com/storage/byodb

Four core components:

• Storage: Table Storage (Snowflake/BigQuery) and File Storage (S3, Azure Blob, GCS) with 7-day snapshots
• Transformations: SQL, Python, R, Julia, native dbt with input/output mapping and version control
• Flows: Visual orchestration with parallelization, conditional logic, CRON scheduling, event triggers
• Components: 700+ extractors, writers, applications running as isolated Docker containers

Connector categories: Databases (MySQL, PostgreSQL, MongoDB), CRM (Salesforce, HubSpot), Marketing (Google Ads, Facebook, TikTok), ERP (NetSuite, SAP), BI destinations (Tableau, Looker, GoodData). Generic connectors extend to any REST API.

Learn more at keboola.com/product/integrations

Keboola offers a consumption-based pricing model combining subscription licensing with usage billing.

Free Tier ($0):

• Unlimited ETL/ELT pipelines, all 700+ connectors, SQL and Python transformations
• 120 minutes in first month, 60 minutes monthly thereafter
• Top-up available at $0.14 per minute

Enterprise Tier:

• Data Catalog and sharing, Data Applications
• Dev/Prod mode with Git CI/CD, VPC deployments
• SOC 2 Type II/GDPR/HIPAA compliance, SAML and SSO
• Dedicated Technical Account Manager

Data remains intact if credits run out—jobs pause but nothing is deleted.

Learn more at keboola.com/pricing

Keboola serves three primary audiences:

• Data engineers build and manage scalable data pipelines without infrastructure overhead
• Data analysts and scientists leverage SQL and Python transformations plus interactive workspaces for analysis
• CxOs and business leaders eliminate shadow IT/AI, consolidate fragmented data tools, and reduce time-to-insight

AI and vibe coders—developers building with AI-first tools like Claude, Cursor, or Windsurf—can control Keboola entirely through natural language.

Common use cases include customer 360 analytics, financial consolidation, marketing attribution, operational reporting, and machine learning pipelines across financial services, retail, marketing agencies, and technology companies.

Learn more at help.keboola.com/overview

Keboola's differentiation centers on end-to-end platform scope versus point solutions:

• vs. Fivetran: While Fivetran handles data ingestion only, Keboola covers ingestion, transformation, orchestration, governance, and AI automation in one unified environment
• vs. Snowflake: While Snowflake focuses on data storage and compute, Keboola integrates with Snowflake (and BigQuery) as storage backends while adding the full operational layer

AI-native architecture sets Keboola apart—the MCP Server enables AI agents to build and manage pipelines via natural language. Multi-cloud flexibility offers deployment across AWS, Azure, and GCP without lock-in. Built-in governance includes automatic compliance monitoring and policy enforcement natively.

Learn more at keboola.com/product/overview

You likely don't need to replace Power BI or Tableau — the real question is what prepares the numbers before they reach the dashboard. BI tools visualise data beautifully, but they don't solve consolidation, governance or financial logic.

Most of the hard work happens before the dashboard:

• Extracting and aligning data from multiple ERPs
• Mapping different CoAs into a group standard
• Translating currencies accurately
• Identifying and eliminating intercompany transactions
• Performing reconciliation and data quality checks

When this logic lives in Excel or custom scripts, you inherit: no audit trail, key-person dependency, manual errors and rework, slow change management.

Keboola provides the trust layer beneath your BI: automated consolidation and governance, certified datasets for reporting, drill-down to journals and source transactions. BI tools continue to do what they do best: visualisation.

The result: Power BI/Tableau dashboards your CFO can trust — because the data underneath is correct, governed and fully auditable.

Technically, yes — but the question is whether building a full finance data foundation is the best use of your team's time. Platforms like Fabric and Databricks are powerful toolkits, while Keboola provides a purpose-built, ready-to-use financial consolidation solution.

Building this internally usually requires:

• Custom ERP/API connectors
• A financial data model with proper hierarchies
• A chart of accounts mapping UI and governance
• FX translation rules with full configurability
• Intercompany matching and elimination logic
• Versioning, QA, reconciliation and completeness checks
• Lineage, auditing and SOX-style controls

This typically means 6–12 months of engineering work before finance sees value — and ongoing maintenance forever.

Keboola accelerates this dramatically: out-of-the-box connectors and financial data model, consolidation rules, mapping logic and governance baked in, end-to-end automation delivered within weeks, and clean integration with Fabric/Databricks if you already use them.

Our philosophy: Build where it differentiates your business. Buy where the problem is already solved.

Keboola plays a fundamentally different role in the finance stack — we solve the data foundation, while EPM tools solve planning and modelling. EPM platforms assume your financial data is already consolidated, harmonised and trusted; Keboola is the layer that makes that data clean, governed and ready for planning.

• EPM tools excel at budgeting, forecasting and scenario modelling
• But they assume the data feeding them is already standardised
• In multi-ERP environments, someone still has to: extract data from every system, map different charts of accounts into one structure, apply FX conversion rules correctly, identify and eliminate intercompany transactions, reconcile and validate before loading
• Keboola automates all of this upstream financial consolidation work
• We can serve as the data backbone for Anaplan/OneStream — or provide built-in planning for simpler needs

In short: EPM tools help you model the future. Keboola makes sure the data from the past and present is accurate enough to trust.

Keboola offers a transparent subscription model designed to scale with your data volume and operational complexity. Pricing reflects the full Financial Intelligence solution — from integration to governance — without locking you into rigid licensing structures.

• Keboola Financial Intelligence is offered as a subscription, with tiers that scale by data volume, complexity and deployment model (hosted vs. your cloud)
• The license includes the full Financial Intelligence solution: data integration, consolidation logic, planning capabilities, AI features, and ongoing platform updates & support
• Most customers see a clear business case based on time saved in consolidation and reporting, reduced tooling and infrastructure complexity, and the ability to redeploy finance capacity to higher-value work

For a tailored estimate and ROI model based on your entities, systems and team size, it's best to talk directly with our team.

Yes – and that's often the recommended approach.

• Start with 1–2 high-impact use cases (e.g. group financial results + daily sales), connect your core systems, and prove value in a single quarter
• Use hard numbers (days to close, hours/MDs saved, error reduction) to build the internal business case
• Then roll out additional entities, use cases (profitability, treasury, risk) and departments on the same finance data platform

A standard rollout follows five phases and typically delivers first value in ~8 weeks:

1. Preparation – environment setup, access, goal alignment
2. Data Processing – connect 2–3 main source systems, initial model, quality checks
3. Governance – business glossary, access model, report catalogue
4. Customisation – KPIs, eliminations, financial model adjustments
5. Enablement – training for finance users, self-service BI setup

Simpler setups (fewer systems, fewer entities) can go live faster; complex multi-entity groups usually land in the 8–12 week range.

AI is there to augment, not replace, finance:

• An AI assistant helps build and manage data pipelines and automations directly inside Keboola – speeding up development for data & finance power users
• An AI chat interface lets business users explore governed data in natural language: ask questions, request ad-hoc cuts, generate draft dashboards – all against certified datasets
• All AI usage sits on top of the same governed, auditable layer, so outputs remain explainable and compliant for finance use cases

Data quality is built into the Financial Intelligence workflow through validation, completeness checks and continuous monitoring. Keboola ensures that all inputs, transformations and outputs remain consistent across systems and entities.

• A dedicated data quality layer validates completeness, consistency and reconciliations across entities and systems
• You can define business rules, thresholds and alerts (e.g. intercompany mismatches, balance discrepancies, missing records) and monitor them in real time
• If you already use data quality tools, they can be integrated; Keboola becomes the central place where clean, trusted financial data is assembled for reporting and AI

Yes — Keboola treats budgets, actuals and forecast versions as equal citizens in the data model. This allows real-time, side-by-side comparison and seamless integration of planning updates into the consolidated view.

• Financial Intelligence supports multiple data versions – actuals, budget, reforecast, multiple scenarios – through the same pipeline
• A planning front-end with a spreadsheet-like interface lets users input or adjust budgets and forecasts with a full audit trail
• Updated forecasts automatically flow back into consolidated views, including currency conversion and roll-ups, so you can run rolling forecasts and scenario analysis without manual rebuilding

The platform is designed to make every data movement, transformation and metric definition fully transparent. You always retain ownership of your data and models, and Keboola ensures all logic remains portable.

• Every object – tables, pipelines, transformations, metric definitions – is versioned and logged
• Visual lineage lets finance teams trace any metric back through the pipeline to the original ERP or source system
• A built-in "Data Takeout" mechanism exports data, transformation scripts and configuration logic to your storage of choice, so you are never locked in

The platform automates group consolidation by standardising journal-level data from each entity into a unified model. It is built to accommodate variations in ERP structures, local currencies and fiscal calendars without manual workarounds.

• Extracts journal entries from each ERP, standardises them into a common structure, and generates group P&L/BS/CF with proper hierarchies
• Handles multi-currency by applying configurable FX rules (spot, average, custom rate tables) at transaction level
• Supports entities with different fiscal year-ends; data is stored at monthly grain and aligned to your group reporting calendar for YTD and prior-period comparisons
• Intercompany transactions are tagged at journal level and eliminated or reported according to your group policy, with transparent, adjustable logic

Full control. Keboola gives finance teams full visibility and authority over how metrics are defined, calculated and adjusted. Nothing is hidden — every rule, mapping and transformation is accessible and modifiable.

• Financial Intelligence ships with a standardised financial model and KPIs, but nothing is a black box
• You can inspect, modify and extend any calculation – from intercompany elimination rules to profitability metrics and allocation logic
• Chart of accounts mapping is supported by a CoA management app and smart suggestions; once mapped, new accounts are easy to classify

Keboola supports audit workflows by enabling controlled, transparent access to governed datasets. External users can be granted scoped permissions while maintaining complete traceability of how every number is produced.

• You can create dedicated auditor roles or workspaces with read-only, scoped access to the data and reports they need
• All transformations and adjustments have a clear audit trail – what changed, who changed it, and when
• This typically shortens audit cycles and shifts conversations from "where did this number come from?" to "what does it mean?"

Keboola is used by regulated financial institutions and is designed for strict governance:

• Enterprise-grade encryption, role-based access control, and full audit logging
• Support for GDPR and data residency requirements (EU/US regions, private-tenant options)
• Architecture and workflows that support SOX-style controls and IFRS/GAAP-compliant reporting processes, with complete traceability from transaction to report

Learn more at our Trust Center.

Customers typically realise tangible improvements within weeks because Keboola automates the high-effort, low-value parts of the finance workflow. The impact is measurable across close cycles, manual workload, data accuracy and the speed of decision-making:

• Close time cut by up to 58% (e.g. 12 days → 5 days)
• Manual work in finance reduced by up to ~70% of team capacity
• 100+ man-days per month saved in group reporting at a global consumer finance group
• 40+ financial data products (P&L, BS, financial performance, portfolio quality, liquidity, profitability) used across CFO, Risk, Sales and Treasury

See our Customer Stories.

We tackle the "Monday-morning nightmare" of modern finance. Finance teams use Keboola to eliminate the operational drag caused by manual work, inconsistent data and slow consolidation. The platform replaces spreadsheets and system silos with a governed backbone that automates the recurring pain points every Finance Leader recognises:

• Slow close cycles (10–25 days) and delayed board reporting
• Up to 70%+ of analyst time spent on data gathering and reconciliation instead of analysis
• Conflicting numbers from different systems and spreadsheets ("multiple versions of the truth")
• Fragile Excel models, key-person risk, and painful audits

Keboola provides hundreds of pre-built connectors for:

• All Major ERPs (SAP, Oracle, NetSuite, Sage, Microsoft Dynamics and others)
• CRM, billing, risk, collections, HR and operational systems
• Databases and cloud storage, plus flat files like CSV/Excel (SharePoint / OneDrive, Google etc.)

And if you have a niche or legacy system, we can connect it via API or file-based integration just as easily.

Keboola Financial Intelligence™ is a solution for Finance teams that connects all your financial systems (ERPs, CRMs, Excel, data warehouses), automates consolidation and reporting, and gives finance leaders a real-time, audit-ready view of performance across entities.

It's built for:

• Mid-market and small-enterprise organisations with 5–50 finance FTEs
• Multi-entity (Groups) multi-system environments with Excel-heavy workflows and no modern finance data backbone

Keboola is designed to fit into your current finance tech stack without requiring migrations or process overhauls. It connects to your existing systems, ingests data as it is today, and unifies everything into a governed model on top of your warehouse.

• Works with your existing ERPs, GLs, CRMs, DWHs and spreadsheets – no re-platforming required
• Ingests data in existing formats (SQL, APIs, CSV, Excel, files) and normalises it into a common financial data model
• You keep your current tools; Keboola sits underneath as the data platform that makes everything consistent and automated

Yes — Keboola is intentionally built as a "bring your own tools" platform. It leverages your warehouse (if you wish to do so) as the storage and processing layer and feeds governed data into the BI tool you already rely on.

• Keboola runs on top of modern cloud data warehouses like Snowflake or BigQuery – either in a Keboola-hosted environment or in your own cloud account
• Data is then exposed to any standard BI tool (Power BI, Tableau, Looker, ThoughtSpot, Excel, etc.)
• You keep your existing analytics front-end; Keboola simply ensures the data feeding it is unified, governed and always up to date

Keboola provides comprehensive audit capabilities:

• Complete audit trails: All user actions, data changes, job executions, and configuration modifications are logged with timestamps and user attribution
• Telemetry data: Job execution details, data flows, schema evolution, and operational metadata available for compliance reporting
• Controlled support access: Keboola support engineers must request access through the platform; project administrators receive notifications and can approve or reject requests
• Time-limited access: When approved, support access is granted with full audit logging and configurable auto-join policies
• Customer-controlled: Organizations can disable auto-join to require explicit invitations for all access

Learn more at keboola.com/product/security

Keboola provides multi-region deployment options across three major cloud providers:

Available regions:

• United States: AWS US Virginia, GCP US Virginia
• European Union: AWS EU Frankfurt, Azure EU Ireland, GCP EU Frankfurt
• Custom regions: Single-tenant deployments in any region supported by the cloud provider

For EU data protection, sub-processors (AWS EMEA SARL, Microsoft Ireland Operations, Google Cloud EMEA Ltd.) are contractually bound to EU processing.

BYODB deployments give customers direct control over data residency since data resides in customer-owned databases.

Learn more at keboola.com/dpa | security.keboola.com

Keboola implements multi-layered access controls:

• Role-Based Access Control (RBAC): Granular bucket-level permissions with roles including Share, Admin, Guest, Developer, and Reviewer
• Single Sign-On (SSO): SAML integration with Active Directory, Azure AD, Google Authentication, and other providers
• Multi-Factor Authentication (MFA): Authenticator apps (TOTP) and hardware security keys (FIDO/U2F)—administrators can enforce MFA for all users
• Token-based authorization: Short-lived tokens with specific scopes for automated processes, restricted to specific buckets or operations
• Brute force protection: CAPTCHA verification after 10 failed login attempts within 5 minutes

Learn more at keboola.com/product/security

Keboola's security architecture adapts to three deployment models:

• Multi-tenant SaaS: Network isolation between tenants, encrypted storage, access controls. Available in AWS (US, EU), Azure (EU), and GCP (US, EU). VPC deployment options available for Enterprise.
• Single-tenant deployment: Complete Keboola stack within customer's cloud environment—maximum control, custom security policies, custom domains, direct Active Directory integration
• BYODB (Bring Your Own Database): Customer data stays in customer-owned Snowflake or BigQuery instances. Keboola orchestrates but never stores customer data on Keboola infrastructure.

All deployment models maintain full audit logging and Docker containerization ensuring components run in isolated environments.

Learn more at help.keboola.com/storage/byodb

Keboola implements defense-in-depth encryption at multiple layers:

Data at rest:

• AES-256 encryption through AWS KMS, Azure Key Vault, and Google Cloud KMS
• All stored data—Snowflake tables, file storage, configuration values—remains encrypted

Data in transit:

• TLS 1.2+ (HTTPS) for all API communications, web interface access, and component data transfers

Application-level encryption:

• Sensitive configuration values prefixed with # are automatically encrypted before storage
• No decryption API exists for end users—encrypted values only decrypted during component execution
• Ciphers are region-locked and cannot transfer between deployments

Learn more at developers.keboola.com/overview/encryption

Keboola maintains enterprise-grade security certifications:

• SOC 2 Type II: Annual certification demonstrating ongoing operational effectiveness of security controls—available through the Trust Center
• GDPR Compliance: Maintained since May 2018 with comprehensive Data Processing Agreement (DPA) and Standard Contractual Clauses
• HIPAA Compliance: Available for healthcare organizations in the Enterprise tier

Cloud infrastructure partners (AWS, Azure, GCP) maintain ISO 27001, CSA STAR, and other certifications. Sub-processors are contractually bound to process data only in approved regions.

The Trust Center provides access to compliance documentation, vulnerability disclosure policies, and bug bounty program.

Learn more at security.keboola.com

Beyond Data Agent and MCP Server, Keboola includes several AI-powered productivity features:

• AI Flow Builder: Creates 3-step ETL flows from user intent—describe what you want and AI suggests the complete Flow structure with data sources and destinations
• AI Component Suggestions: Intelligent recommendations when searching for connectors, enabled via Project Settings
• AI-Generated Descriptions: Automatically documents configurations, transformations, components, and flows—addressing documentation debt
• AI Error Explanations: Translates cryptic job failures into human-readable explanations with suggested resolutions

These features accelerate development while maintaining full auditability. Users can customize AI behavior through AI Rules to match organizational standards.

Learn more at help.keboola.com/ai

Keboola provides end-to-end AI/ML infrastructure:

• Generative AI integration: Query AI models (OpenAI, Azure OpenAI, Google Gemini) with Keboola data. Includes API key management, custom prompts, and cost protection.
• Embedding providers for RAG: OpenAI, Azure, Cohere, Hugging Face, Google Vertex AI, AWS Bedrock. Pre-built Data Templates accelerate RAG development.
• MLflow model deployment: Deploy trained models directly in Keboola with unique endpoint URLs for inference.
• ML-ready data preparation: Python, R, and Julia transformations for feature engineering and preprocessing.
• Development flexibility: Keboola Workspaces (Jupyter-style), AI agents (Claude, Cursor), or traditional code (SQL, Python, dbt) with Git CI/CD.

Learn more at help.keboola.com/ai

Keboola addresses AI governance through platform-level controls:

• AI Rules: Administrators define custom instructions for AI behavior—global or component-specific. Control communication style, response format, language, and focus areas.
• Complete audit trails: All AI interactions are logged—who requested, what context provided, what proposed, whether approved.
• Approval workflows: AI never autonomously changes production. Changes are proposed, reviewed, and only shipped after explicit approval.
• Data isolation: Multi-project architecture restricts AI access to specific projects.
• Privacy by design: Privately deployed Azure OpenAI—prompts and responses are NOT used to train the model.
• Cost protection: Credit/token limits per run with real-time cost estimates.

Learn more at help.keboola.com/management/project

Keboola Data Agent is a context-aware AI assistant embedded in the platform that enables users to build, debug, document, and manage data pipelines using natural language.

How it works: Data Agent runs within projects with full access to metadata and governance controls. It uses the platform's semantic layer to ground responses in actual project context—ensuring suggestions are deterministic, traceable, and reproducible.

Capabilities:

• For data engineers: Generates SQL transformations, dimensional models, and integration code from prompts. Debugs failed jobs and auto-generates documentation.
• For data analysts: Queries data conversationally, returns instant charts and tables, writes and executes SQL with full transparency.

All suggestions require explicit user approval before execution, and every action is logged.

Learn more at keboola.com/product/agent

Keboola enables conversational interaction with the entire data platform—not just querying data, but building pipelines, debugging failures, and generating documentation through plain English.

Examples:

• Data querying: "Show me revenue by product line this quarter vs. last quarter"
• Pipeline creation: "Build a pipeline to extract daily sales from Shopify, join with CRM data, calculate lifetime value, load to Snowflake. Schedule daily at 6 AM."
• Debugging: "The transformation in Flow_X failed. Show the failing query, explain the issue, propose a fix."
• Documentation: "Create full project documentation for all flows in markdown format."

Every interaction produces auditable, reproducible outputs—AI proposes, user approves, actions are logged.

Learn more at keboola.com/product/agent

The MCP Server (Model Context Protocol) is an open-source bridge that exposes the entire Keboola platform as callable tools for AI agents.

Supported AI tools:

• Claude (Desktop and Claude.ai)
• ChatGPT (Plus/Pro)
• Cursor, Windsurf, VS Code with GitHub Copilot
• Workflow automation tools like MAKE

Capabilities through MCP:

• Search and query data directly
• Set up complete ETL workflows
• Build and test SQL transformations
• Launch and monitor pipeline jobs
• Auto-document projects down to column level

All features use privately deployed Azure OpenAI—prompts and responses are NOT used to train the model.

Learn more at keboola.com/mcp