top of page

Causal AI

Eradicating data misinterpretation

Conventional data analytics and AI methods are by design limited to understand correlations only.


Causal AI takes a fundamentally different approach: It goes beyond correlations by learning the causal structure underlying your data.

A correct understanding of cause-effect relations is crucial for informed business decisions, as highlighted by the example below.

Conclusions from correlations

  • Playing more golf will increase life expectancy.

  • Golfing activity is a good predictor for life expectancy.

Conclusions from causal model

  • Playing golf does not increase life expectancy. Instead a higher income increases both life expectancy and golfing activity

  • Golfing activity is a not a robust predictor for life expectancy. Instead one should use income.

Why Causal AI

Discrimination-free insurance pricing

Ensuring that insurance pricing is free from discrimination against protected characteristics such as ethnicity.

Root cause analysis

Discovering the causes of process faults and inefficiencies.

Customer churn

Finding the right treatment for the right customer at the right time.

Improved loan decision making

Assisting lenders to make data-driven decisions for loans nearing default.

Drivers of demand

Building a causal model to identify the drivers of demand for a product.

Case studies

Your enterprise data

AI refinement

Data refinement is a major part of every data analytics project. We leverage AI to accelerate the process of improving the quality and usability of raw data. 

Cleaning & integration

We use AI to combine datasets with different formats from multiple sources and clean them.


AI allows the quantification of unstructured datasets (text, images, etc.) with low effort.


We enrich datasets with additional information (e.g. using LLM powered web research agents).

Causal model

Step 1: Using algorithmic methods paired with human expert knowledge, we first create a DAG (directed acyclic graph) to represent the causal structure underlying your data.

Step 2: Based on the DAG, we design a Causal AI and train it with your data.


We query the Causal AI model to answer your questions and help you design the best plan of action.


Find the root cause of some observed effect in your data.


Create robust predictions based on causal structures instead of correlations.


Run through hypothetical scenarios, such as: What happens to X if we change Y?


Find the optimal action for a given objective.


Understand the “why” behind the model’s recommendations, instead of relying on black-box forecasts.

Our approach

Big Tech is betting on causality

Recently, there has been an exponential increase in research activity and commercial application of causal AI in the tech industry.

Causality in industry

bottom of page