AI Data Foundation

Your company data,ready for AI

Give AI access to the data and context it needs for smart, high-quality answers.

Start free trial Get a demo

Your AI, connected with the AI Data Skill, and your data apps read through Koalake. Everything from the context store (definitions, memories and data from your systems or your data platform) passes permission management first, which sets who sees what.

AI Data Skill

Connect your LLM

Data App Platform

One-click share Data write-back

Your systems or your data platform

Give controlled access to data

When you give AI access to your data, it’s wise to manage permissions carefully. With the Koalake permissions engine, you stay in control of who exactly gets access to what.

One access rule for Sales: only the customers in their own country, and all data linked to them. Asked which customers are more than 60 days overdue, Tom’s AI (Sales NL) finds 4, Jonas’s AI (Sales DE) 5, and Nadia’s AI (Finance) all 13.

Each person sees only the data they need.

One meaning of “revenue” for every AI

Ask two people’s AI for revenue and you can get two figures, depending on how the AI decides to calculate it. A semantic model records how a metric is calculated, so each AI gives the right answer.

A metric definition for revenue: from invoices, not won deals; counted on the invoice date; after credit notes. Each rule points at the column in your ERP tables that carries it, so every AI counts revenue the same way.

One clear definition of how revenue is calculated.

A knowledge base that grows naturally

As people work, their AI keeps notes on what we learn. Shared memories are linked to the data they relate to and are protected by access rights, just like your data.

Revenue invoiced per month in 2026, with two notes: March includes a one-off credit note of €180,000, the radiator recall. Leave it out of trends. August was lower because of high sick leave that month.

Put your data into context with memories.

Expose high-quality datasets

Connect your data to Koalake and expose clean datasets for your people and AI to use.

Do you have a data platform?

ProBring your data together in Koalake.

With Pro, Koalake becomes your data platform, with the data from all your systems in one place. Build your own pipelines in Airflow and transformations in dbt. Our team helps you with the setup.

Talk to sales

BasicKoalake connects to your existing infrastructure.

With a data platform on AWS, Azure, Google Cloud, Snowflake or Databricks, our team connects Koalake to it during onboarding. Your team’s models and tables stay as they are, ready for your AI to read.

Included in every edition

Your platform not listed? Talk to us

Built on widely used open-source tools

The AI Data Foundation runs on open-source tools that many data teams already use.

DuckDB
Runs every query your AI and data apps send: a fast, open-source query engine.
dbtIn Pro
Holds each calculation, such as revenue after credit notes, as a readable file.
AirflowIn Pro
Runs each of those steps on its schedule.

Need a data platform? Go Pro.

Basic extends the platform you already have.
Pro gives you a fully functional data platform.

You have a data platform

Basic

Included in every edition

  • Connects to your data platform
  • Permission management, down to the row
  • Context store: definitions, memories and data

Your data is in your systems

Pro

Talk to sales

  • Everything in Basic
  • Your systems connected directly, no data platform needed
  • Clean datasets, refreshed on a schedule
  • Transform your data with dbt
  • Scheduled processes with Airflow

Questions & answers

We already have a data platform. Do we need Pro?

No. Basic extends your current data platform. Pro is for customers that want to use Koalake as a fully functional data platform.

Does Koalake change anything in our systems?

No. By default Koalake’s connections only read. What people type into a data app is saved in Koalake, not written back to your systems, unless you set up a process to do so.

Can a prompt get around someone’s role?

No. Your AI connects to Koalake as you, under your role, and Koalake checks access where the data lives. A prompt can change what your AI asks for, but not what Koalake returns.

What happens when someone moves to another team?

IT changes their role in Koalake. Their AI and every data app they open follow it, because every request is checked against their role. Nobody has to update each system.

Could our own data team build this?

Yes, and with a data platform they may have most of it. Basic adds one set of roles for people, their AI and data apps on top. Without a data platform, Pro gives them one, set up by our team during onboarding.

Where is our data stored?

In the EU, by default: Koalake stores and processes your data there. Which cloud it runs on, and what your AI provider receives, is on our security page.
How we keep your data safe

Start on demo data today