Learning Advisor Engine

Intro

The Personalised Learning Advisor is the learner-facing experience.

Behind it sits the Proof Harbour Orchestration Engine, a workflow layer built to guide a user through a structured sequence of steps while maintaining session continuity, controlled progression, and connected outputs.

This page explains the engine at a high level. It shows the problem being addressed, the design approach behind the system, and why structured orchestration matters when AI is used in settings where evidence, consistency, and accountability count.

A unique session ID is created at the start of the journey. The learner can copy or download that ID and return within 3 months to continue from where they left off.

The problem

Many AI tools work as open-ended conversation. That suits exploration, but it is weaker when the goal is to support a defined process with clear stages, reliable outputs, and the ability to return to the same journey later.

In practice, this can create issues such as outputs drifting apart, weak continuity when a user returns, resources appearing before they are ready, and poor visibility over what is complete, pending, or still being prepared.

The Learning Advisor Engine was built to address those issues through a staged, resumable, and controlled workflow model.

The approach

The Proof Harbour Orchestration Engine is built on a simple principle.

A learning journey should not depend on a loose exchange alone. It should be supported by a workflow engine that manages state, progression, output handling, and readiness checks behind the scenes.

• Maintains a server-side session record.
• Moves the learner through defined stages.
• Generates linked outputs from the same session context.
• Validates asynchronous resource generation before the journey moves forward.
• Supports return and continuation through a session identifier.
• Allows the learner to copy or download the session ID for later return.
• Keeps the visible experience simple while the orchestration layer manages the structure underneath.

How the journey works

The live Personalised Learning Advisor currently follows a staged workflow.

• Interview – The learner sets out what they want to learn.
• Summary – The system creates a structured view of the learner’s direction.
• Learning Plan – A plan is generated from that structured context.
• Resources – Relevant materials are prepared and presented in a controlled way.
• Review – The learner reflects on progress, captures what has been achieved, and identifies next actions.
• Next Steps – The journey produces a forward path rather than ending at a single output.

The learner sees a guided journey. The orchestration engine manages the state, dependencies, and continuity between those stages.

What sits behind the interface

The visible interface is only one part of the system.

The engine behind it is designed to support resumable staged sessions, controlled progression between steps, linked outputs, validated asynchronous resource generation, and session-based return and continuation.

This includes session-based return through a unique ID issued at the start of the journey and used to resume progress later.

This matters because a learning journey becomes more useful when the outputs relate to each other and can be revisited with continuity.

Resource readiness

One of the important design choices in the engine is controlled handling of the resource step.

Rather than treating resources as immediately available, the workflow model is designed around validated asynchronous generation and readiness checks.

This reduces the risk of the learner moving into a stage before the underlying output is properly available.

Why this matters

The value of this approach is not in AI as a label.

It is in using structured workflow design to make digital systems easier to understand, trust, review, evidence, and improve.

That matters in education. It also matters more broadly in any setting where a system needs to do more than generate text. It needs to support a process and leave behind something coherent.

Patent status

The underlying orchestration engine is the subject of a UK patent application filed on 10 March 2026.

Patent Pending – UK patent application filed 10 March 2026

The filing relates to the technical architecture behind a resumable staged AI workflow, including structured session handling, validated asynchronous resource generation, and linked output management.

Live example

A live example of the learner-facing implementation is available here.

Naming

Personalised Learning Advisor is the learner-facing label.

Proof Harbour Orchestration Engine is the underlying technology label.

This distinction separates the user experience from the workflow engine that powers it.