VIRTICUSRequest Assessment
Back to the wider record

AI Governance

Governing AI-Driven Learning Recommendations

AI-personalised learning paths that had to be suitable, explainable, and not misleading.

Illustrative engagement

The engagement

Where the decision had to hold under scrutiny

AI was used to personalise learning pathways and generate recommendations for students. The main challenge was ensuring that outputs were suitable, explainable, and not misleading when used in real educational settings. Control was introduced around how recommendations were reviewed and applied. This helped keep learning decisions appropriate, consistent, and defensible.

Key challenges

The problem, and what was done about it

Challenge

AI was personalising learning pathways and generating recommendations for students, where the outputs had to be suitable, explainable, and not misleading when applied in real educational settings.

What was done

Introduced control over how recommendations were reviewed and applied, keeping learning decisions appropriate, consistent, and defensible rather than accepted on trust.

ReasonAssure

Illustrative — reconstructed from the engagement context

How Verdica makes it defensible

The same five stages, applied to this decision

Every engagement runs the same path. For this one, the weight falls on Reason, Assure — the stages where a decision of this kind is most often challenged.

01

Trace

Map every point where the decision could be challenged — before a line of logic is written.

02Core

Reason

Build the logic in the open: explainable, deterministic, same input, same output — no black box.

03Core

Assure

Validate coverage and oversight before anything ships; nothing unassessed is waved through as safe.

04

Certify

Lock an immutable, versioned record of exactly what was decided and on what basis.

05

Evidence

Hand over the audit trail — the pack that holds the moment someone says prove it.

Discuss a related situation

A decision like this can be made defensible — on the record, not after the fact

A short discussion is usually enough to locate where the risk sits — in the logic, the controls, the evidence, or a combination — and where Verdica makes it hold.