Our methodology · Machine in the Loop

Educational intelligence
Human judgement

A shared data foundation. A model built with experts. A way to examine what could change.

Ask

Begin with the institution’s question

Which conditions could improve student progression? What should change in a course? Where could support make a difference? Educators and institutional experts define the question, the goals and what counts as a meaningful outcome.

The machine enters their process to help make that reasoning explicit. Educational decisions remain with the people responsible for them.

Model

Build the model together

A common data integration layer connects institutional sources. The proprietary causal model is co-built with domain experts and educational scientists, making the hypothesised relationships between context, interventions, mechanisms and outcomes explicit.

Expert knowledge frames the assumptions. Research and institutional evidence help examine and refine them.

Open a technical example of the causal model
Context CISEEIntervention IGuidance during studiesSatisfactionMotivationPerformanceOutcome YOn-time progressionContext CISEEIntervention IGuidance during studiesSatisfactionMotivationPerformanceOutcome YOn-time progression
Technical example supplied by Didaflow. Arrows represent proposed causal relationships, not effect estimates. The model is co-built with domain experts and educational scientists.

Explore

Compare possibilities before acting

What-if scenarios help an institution explore an intervention before committing to it. The platform brings programme monitoring and scenario exploration into the same environment.

Estimates depend on the model, its assumptions and the available evidence. They inform judgement and should remain open to challenge.

Inspect

Keep sources and uncertainty visible

The methodology calls for every assumption to be recorded and every effect estimate to carry its source: research, institutional history or declared expert judgement. Uncertainty belongs in the reasoning.

When evidence is insufficient or sources disagree, the expert should see that limitation and decide how to proceed.

Learn

Make revision part of the work

The long-term aim is a learning institution: decisions designed from explicit hypotheses, evaluated against declared goals and used to enrich shared institutional knowledge.

Educational decisions belong to educators.

Read the full white paper for the principles, research origins and responsibilities behind Machine in the Loop.

Read the white paper

A little curiosity

The decorative figures in the opening scene and stakeholder sections come from Children’s Games by Pieter Bruegel the Elder. Their gentle movements accompany the page as you scroll.