AI Report Section Drafting

Overview

AI Report Section Drafting generates first drafts of report sections using your project's actual observation data. Instead of starting from a blank page, you get a structured draft that you can review, edit, and refine.

Supported section types:

• Survey Methodology -- summarises the survey design, methods used, and effort applied
• Species Summary -- compiles species records, conservation status, and distribution across surveys
• Observation Summary -- aggregates observation counts, spatial patterns, and temporal trends
• OHS Compliance -- summarises workplace health and safety incidents, field conditions, and compliance records

AI drafting is powered by Azure OpenAI. All data stays within your tenant boundary and is never used to train external models.

How to Use

1. Open a report and navigate to a section that supports AI drafting. Supported sections display an AI Draft button with a sparkle icon in the section editor.

2. Click AI Draft to load a preview of the data that will be sent to the AI.

3. Review the Data Context Preview, which shows:

   • The number of observations included
   • The number of species referenced
   • The number of surveys covered
   • Any observations that were excluded due to cultural sensitivity

4. If everything looks correct, click Generate Draft to start generation. This uses one draft from your monthly quota.

5. Watch the draft appear in real time as it streams. You can click Cancel at any time to stop generation.

6. Once generation is complete, review the text carefully, then choose one of three actions:

   • Accept -- inserts the draft into your report section for further editing
   • Regenerate -- discards the current draft and generates a new one (uses another quota credit)
   • Discard -- cancels without inserting anything

AI-Assisted Draft Label

Sections that contain AI-generated content are permanently labelled with an AI-assisted draft badge. This label appears in the section header and cannot be removed.

The label ensures transparency for regulatory submissions and internal review. Reviewers and approvers can always see which sections originated from AI-generated content.

You can freely edit the content after accepting a draft. The label indicates the section's origin, not its current state -- the text you submit may be entirely rewritten.

Monthly Usage Quota

AI drafting is limited per month based on your subscription plan. The current usage count is displayed next to the AI Draft button (e.g., "5 / 250 AI drafts this month").

When the monthly limit is reached:

• The AI Draft button is disabled
• A message explains that the quota has been exhausted
• The quota resets automatically at the start of your next billing period

If you need additional capacity, contact your administrator about upgrading your subscription plan.

Data Privacy and Cultural Sensitivity

Observations flagged as culturally sensitive are automatically excluded from AI processing. This includes observations related to:

• Cultural heritage sites
• Sacred sites and areas of cultural significance
• Burial sites and sensitive archaeological contexts

The number of excluded observations is always shown in the Data Context Preview, so you know exactly what data the AI can and cannot access. This is a mandatory privacy protection that cannot be overridden.

If all observations in a section are culturally sensitive, the AI will report insufficient data and generation will not proceed.

Tips for Reviewing AI Drafts

• Verify species names and conservation status against your raw data. AI may occasionally misattribute or omit species from the summary.
• Check observation counts and survey details to ensure they match your project records.
• Adjust tone and style to match your organisation's reporting standards. AI drafts use a neutral, professional tone that may differ from your house style.
• Add site-specific context that the AI may not have captured, such as local conditions, access constraints, or stakeholder considerations.
• Cross-reference with auto-generated sections. Species lists and observation tables in auto-generated sections are pulled directly from the database -- use them as a ground truth when reviewing AI narrative text.
• Remember: AI drafts are a starting point, not a finished product. Every draft should be reviewed by a qualified professional before submission.

Related Articles

• Authoring and collaborating on reports -- Full guide to the report editor, review workflow, and version history
• Auto-generated content types -- How data-driven sections work