Organise the evidence under each pleaded issue, with confidence scores and the places the evidence is thin.

Issue analysis

Organise the evidence under each pleaded issue, with confidence scores and the places the evidence is thin.

InsightsTemplatesIssue analysis

What this template produces

One section per issue tag, each summarising what the evidence establishes, listing the strongest supporting documents with their confidence scores, and naming where the evidence is thin or contradictory.

What it reads

This Insight reads an issue tag and confidence score per document — the output of the Relevance and Issues prompt. If your case has not run it yet, Claira offers to run it before generating.

Claira IssuesMemoissue tags (ISS1, ISS2…) with a confidence score

Two ways to reach it

You do not have to remember this template exists. If you run the Relevance and Issues prompt as a bulk task, the completed task offers this Insight directly:

On a finished bulk task: “These results can make a Issue analysis.” Select Generate insight and you are done — the source is already known.

Otherwise, open Insights and select Issue analysis from the template row at the top of the New insight card. Claira looks for a completed task that ran the prompt and tells you what it found.

The insight prompt

This is what the template puts in What do you want?. Edit it freely — once it is in the box it behaves like any other insight prompt.

Organise the tagged documents under each issue. For every issue, summarise what the evidence establishes, list the strongest supporting documents with their confidence scores, and note where the evidence is thin or contradictory.

Step-by-step in Claira

Open Insights in your case and select Issue analysis from the template row.
If a completed task already ran the prompt, the source fills in and you can go straight to the last step.
If nothing has, select Run this prompt. Claira opens Bulk Tasks with the Relevance and Issues prompt filled in — choose your documents and a destination field, then start the scan and check the estimated token usage on the confirmation.
When the task finishes, it offers to generate the Insight. Nothing is queued or charged while it runs.
Edit the prefilled insight prompt if you want, then select Generate Insight.

Already have this data from somewhere else? Select Pick a field instead and choose the field yourself. Claira does not guess which of your fields holds the right kind of value.

The template points at the Relevance and Issues prompt template — the same one you would pick yourself from the prompt library. The insight template does not carry its own copy, so improvements to the prompt reach every case.

Recommended customizations

Map the issue tags to their real names in the insight prompt so the output reads like your pleadings, not like ISS1.
Ask for a confidence floor — low-confidence tags can drown a section.
Ask explicitly what is missing for each issue; it is more useful than another list of what you have.

Troubleshooting

One issue swamps the rest. Generate one Insight per issue by narrowing the prompt, rather than asking for all of them at once.
Confidence scores look uniform. The per-document prompt probably lacks case context. Add it and re-run the scan.

Issue analysis

On this page