Turn the method into a shared working standard.

What you will find

A complete view of the system before going deeper.

• Three movements: acquire, retrieve and review.
• A protocol to separate fluency from evidence.
• Prompts, workflows, routines and tools to execute.

Learning assistants

Simple roles for research, practice and review without outsourcing judgment.

10 progress scales

Locate current evidence before choosing the next practice.

• 0–2: recognise, question and map.
• 3–5: explain, apply and compare.
• 6–9: defend, teach, adapt and renew.

Learning modes

One structure can support learning, updating, changing fields or working together.

• Self-learning with personal evidence.
• Upskilling and reskilling tied to a real outcome.
• Collaborative learning with shared sources and criteria.

The fluency illusion

A fluent answer is not necessarily correct or sufficient.

• Request citations and open the source.
• Separate facts, inferences and gaps.
• Test through retrieval, application and contrast.

Three frameworks

Organise what exists, what you can do and how consistently you can do it.

• Body of Knowledge: concepts, relationships and sources.
• Capability model: actions you can perform.
• Maturity model: repeatability, evidence and limits.

Six study techniques

Useful practices when connected to purpose and evidence.

• Retrieval and spaced practice.
• Interleaving and elaboration.
• Dual coding and concrete examples.

Apprehend

Retain enough to retrieve, apply and defend without relying on the prompt.

• Explain without notes and find where understanding breaks.
• Produce a project, pitch, presentation or test.
• Use feedback to correct the mental model.

Deep research with AI

Research begins with the plan, not the tool.

• Declare purpose, scope, tensions and criteria.
• Prioritise primary sources and real diversity.
• Synthesize with citations, residual risk and the next decision.

Integrity protocol

Trust comes from tracing what supports every claim.

• Do not invent data or citations.
• Do not treat model agreement as independent evidence.
• Declare unknowns and stop when authority is missing.

(R)Evolve

Reviewing what you learned also means releasing what no longer serves.

• Keep what remains useful.
• Update what is valuable but incomplete.
• Replace or retire obsolete knowledge with a documented reason.

Method tools

Choose the smallest combination that supports research, practice and verification.

• A conversational AI to structure the research plan.
• NotebookLM for working with a source base.
• The open skill to repeat the workflow with guardrails.

NotebookLM

A space for conversing with selected sources and reviewing citations.

• Centralise only sources you are allowed to use.
• Configure a teacher, advisor or coach for the purpose.
• Check responses against sources; the interface does not guarantee accuracy.

High-performance prompts

Fourteen entry points for research, checking, practice and transfer.

Learning workflows

Three indicative paths; time varies by topic and iteration.

• Essential practice: about 30 minutes.
• Ten-step route: 1 to 2 hours.
• Extended route: 2 to 4 hours.

Suggested routines

Cadence sustains the method better than a single session.

• Daily curiosity: record one useful question.
• Weekly apprehension: retrieve and explain without notes.
• Periodic audit: review currency, usefulness and gaps.

Learning standards

Practice ends when reviewable evidence exists, not when content was consumed.

• Method first; tool second.
• Sources before certainty.
• Explain, apply and transfer before declaring progress.

Essential glossary

Ten concepts for a shared language.

• Research plan · verifiable questions, sources and criteria.
• BoK · a map of knowledge and sources.
• Apprehend · retrieve and apply without dependency.
• Triangulation · contrast independent evidence.
• Sufficiency · evidence adequate to purpose and risk.
• Teach-back · explain to reveal understanding.
• Skill · reusable instructions and resources for a capability.
• Prompt · an executable instruction for AI.
• Primary source · evidence close to the fact or authority.
• coverage_gap · a declared gap that prevents closure.

Frequently asked questions

Short answers before you begin.

• Do I need a paid plan? Essential practice can work without one; provider conditions change.
• Does AI learn for me? No. It structures, proposes and supports review; understanding and decisions remain human.
• Can I use any source? Only authorised, relevant and traceable sources.
• When do I stop? When evidence is sufficient for the purpose and residual risk is declared.