Report generation and variants

Phase 10 is the Final Report Workspace. You configure how the report should read, generate the deterministic body, optionally enrich narrative sections with AI, and issue labeled PDF versions. One assessment credit covers the whole lifecycle, with a separate limit on costly AI enrichments.

Quick reference

Action	Counts against credit?	Counts against AI limit (5)?
Change report setup (purpose, audience, verbosity, caps)	No	No
Start / Restart (rebuild tables and structure)	No	No
AI Fill (GPT narratives)	No	Yes — only when settings are new
Issue report (save a labeled PDF)	No*	No
End Assessment (first issue + archive)	Yes — one credit	No
Re-download an issued report	No	No
Clone assessment (fresh AI budget)	Yes — on clone conclude	Resets to 5

*Your credit is consumed when you End Assessment (conclude the assessment). After that you can still change structure and issue more reports without another credit.

Report setup drives structure and tone

In Phase 10 you choose:

• Purpose — board decision, audit evidence, engineering remediation, and others
• Audience and technical depth
• Verbosity (terse, standard, detailed)
• Caps — how many risks, controls, and remediation tasks appear in the main body
• Appendix — include or omit detailed registers

Start / Restart applies these settings to the deterministic report (tables, registers, backend blocks). Restart is unlimited and does not call GPT. Changing setup after a report exists shows a banner: save setup, then Restart so structure matches your preferences. AI Fill alone does not rebuild tables.

AI Fill vs issuing a report

These are separate steps:

1. AI Fill — fills GPT narrative blocks (executive summary, introduction, etc.) using your current setup. This is the costly step.
2. Issue report — renders a PDF and stores it in your issued reports index with a label you provide (for example “Board pack Q2”).

You can issue a deterministic-only report (no AI Fill) or issue after AI Fill. Issuing does not consume an AI enrichment slot.

Five AI enrichments per assessment

Each assessment includes up to five AI enrichment runs — one per unique combination of report settings (a configuration fingerprint).

• If you run AI Fill with settings you already enriched, Riskonami reuses cached text — no GPT call and no slot consumed.
• If you change settings to a new combination, the next AI Fill uses one slot.
• Restart and Issue report are unlimited and do not use slots.

When all five slots are used, you can still:

• Change report setup and Restart
• Issue deterministic reports
• Download any report already in your index

To get new AI narratives with a fresh budget, clone the assessment from your profile. The clone starts with five enrichments again; concluding the clone uses one credit (same billing model as the original).

After you End Assessment

End Assessment (conclude):

• Consumes your assessment credit
• Locks phases 1–9 (assessment data is read-only)
• Archives the assessment for audit
• Keeps the report studio open

You can continue to:

• Adjust Phase 10 report setup
• Restart the report structure (unlimited)
• AI Fill until your enrichment budget is exhausted
• Issue additional labeled reports (unlimited)

Issued reports index and labels

Every time you Issue report, you enter a label describing why that version exists. Issued reports appear in:

• The Final Report Workspace index
• Your profile (alongside the assessment)

Each entry shows:

• Issue number (#1, #2, …)
• Your label
• A short summary of settings (purpose, verbosity, appendix)
• Whether AI enrichment was used
• Download PDF

Use distinct labels so teammates know which file to open later.

Duplicate settings warning

If you select report attributes that match an already issued report, Riskonami warns you before AI Fill or Issue report:

• Download the existing report — same content, no extra cost
• Change setup — different fingerprint, may use an AI slot on Fill
• Issue anyway — create a new labeled copy with the same settings (no new GPT if enrichment is already cached)

This helps avoid accidental duplicate work and unnecessary AI spend.

Recommended workflow

Related docs

• How assessments work — lifecycle and credit timing
• How billing works — credits and reservations
• How sessions work — clone and archive behaviour
• Risk Assessment Walkthrough — UI tour