> For the complete documentation index, see [llms.txt](https://docs.cortex.aic.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cortex.aic.io/workflow-guides/analyse/analytical-techniques.md).

# Analytical Techniques

**Open this area:** In the CORTEX menu, select **Intelligence cycle → Analyse**.

Use a proportionate assumptions check, competing-hypotheses matrix or structured challenge with exact inputs and attributable contributions.

**Who uses this:** Analyst, Reviewer\
**Open in CORTEX:** `/analyse`\
**In-app guide:** `/help?topic=analytical-techniques`

**Expected result:** A reviewer can reconstruct the method version, question, participants, inputs, conclusions, limitations and unresolved independent challenges.

## Before you start

* Create the analytical revision with the exact question, evidence, hypotheses and shared assumptions you intend to examine.
* Use current Analysis and Memory read access; facilitating and interpretations need Analysis write, independent challenge/resolution needs Analysis review and independence from analytical and technique authorship.
* Choose method depth to match the commission, impact, complexity and uncertainty. A light check may suffice when its limitations are explained.

## What to do

{% stepper %}
{% step %}

## Open techniques from the analysis

Choose **Open techniques and challenge** on the analytical detail page.

The session is recorded against the current exact content revision and hash, with your authenticated identity as facilitator.
{% endstep %}

{% step %}

## Select a proportionate method

Explain why the selected technique and challenge depth fit the problem.

* A key-assumptions check records support, review conditions and retained/challenged/retired disposition for selected shared assumptions; you own that register.
* A competing-hypotheses matrix assesses the same selected evidence against at least two hypotheses.
* A structured challenge records an alternative, bias concern, discriminating check and outcome.
  {% endstep %}

{% step %}

## Retain conclusions and limitations

Complete the chosen method, conclusions and limitations, then retain the session.

Matrix interpretation is explicit for every selected pair, including contradictions and discriminating evidence. No ranking, vote or method result changes likelihood or confidence.
{% endstep %}

{% step %}

## Invite attributable participation

Select the retained session.

Participants record their own interpretations and evidence references. Independent reviewers record challenge and bias concerns. The facilitator records the accountable response, and only the original challenger can resolve it after that response.

Unresolved challenges remain visible, including on superseded sessions.
{% endstep %}

{% step %}

## Revisit assumptions without deleting history

Retain a new session and optionally supersede your own latest session of the same method.

Record a revised or retired assumption disposition against current exact inputs. Earlier sessions, support, owners, contributions and analytical uses remain unchanged.
{% endstep %}

{% step %}

## Cite the session and review

Amend the analysis and select up to four retained sessions in a judgement's structured confidence basis.

Saving binds their exact summaries and contributions at that point. New product revisions and ordinary authored/model-assisted exports retain these references. New sessions/contributions request fresh peer review.

Record significant customer-facing dissent through the separate dissent workflow when appropriate.
{% endstep %}
{% endstepper %}

## Check the result

* Does the matrix retain evidence contradicting the favoured explanation?
* Can the reviewer open the exact source revision and identify the actual participants?
* Are limitations and method-selection rationale explicit?
* Are unresolved challenges visible before submission?
* Did changed analytical inputs prompt deliberate reassessment rather than silent refresh?

## If you get stuck

<details>

<summary>Why is the method refused?</summary>

Use populated fields for the selected method only.

Matrices need at least two hypotheses and all selected evidence/hypothesis pairs, with at most 100 cells. Assumption keys and evidence must exist on the pinned revision.

Complete selection rationale, conclusions and limitations.

</details>

<details>

<summary>Why can I not save during review?</summary>

Author content and technique work is locked during peer review.

An independent reviewer must request changes; recording independent challenge also requests fresh review. Authors cannot use a technique save to bypass that lock.

</details>

<details>

<summary>Can a response close a critical challenge?</summary>

No. A facilitator's response leaves the challenge unresolved.

Only its original independent challenger may resolve it after a response, with a retained reason. Earlier statements stay visible.

</details>

<details>

<summary>What if someone else saved or access changed?</summary>

Stale writes and revoked permissions are refused.

Preserve your unsaved wording, reload current state and reconcile deliberately. Failed writes do not discard retained records.

</details>

<details>

<summary>What happens at the limits?</summary>

One analysis retains at most 16 sessions and 64 contributions, plus existing content/audit/storage limits. A judgement cites at most four sessions.

Product output limits include the technique summary; oversized output is refused rather than truncated. Scope a follow-on analysis when necessary.

</details>

## Responsibility and limits

* The facilitator owns the recorded method and assumption register; appointments do not grant permissions.
* Participants are authenticated actors, not an editable attendance list. Independent challenge cannot be overwritten by the author.
* Method facilitators and authored interpretation/response participants remain contributors for independent review and linked product approval. Analysts and reviewers remain responsible for weighing evidence. These records do not replace professional judgement or release authority.

**Reference and scope:** Registered analytical standards, AQuA and red-teaming guidance inform this retained reasoning workflow. Completing a method does not prove objectivity, calculate truth or generate a probability.

**Related workflows:** [Develop and independently review an assessment](broken://pages/79db57a6c6245feef3fb58e4fe6d4a507b2a74a0); [Explain confidence, changed judgements and dissent](broken://pages/b5f23381c9372db8261edb88339424566cabc7f2); [Agree an analytical commission and decision need](file:///7397340/direction/commissioning.md); [Author an intelligence Product and inspect its exports](file:///7397340/product/products.md).