Upload the decision record.
Get a structured assurance review.

Upload the reasons, the file, and any related material for a tribunal decision. The agent reads them, maps the evidence to all 32 assurance checks across 6 principles of good tribunal practice, and produces an auditable draft review for the member to verify and sign off. It is an assurance layer only — it does not make, recommend or pre-judge the decision.

1
Upload
PDF, DOCX, XLSX, CSV, TXT
2
Analyse
The model reads the reasons and the file
3
Map
32 checks, cited evidence
4
Review
Auditable draft for member sign-off
1API Key
Your key is used only in your browser and never stored or sent anywhere except OpenAI's API. Get one at platform.openai.com/api-keys — ensure it has access to the Chat Completions API. Calls are made directly from your browser, so no server is required.
2Project Details
3Upload the decision record
📄
Drop documents here or click to browse
PDF · DOCX · XLSX · CSV · TXT · MD — multiple files supported
6 principles · 32 checks · ~2 min
Assessment Running
Reviewing the record…
Initialising… 0%
AI-Generated Draft · assurance layer only — This review was produced by an AI model from the documents you uploaded. It does not make, recommend or pre-judge the decision; the tribunal member remains the decision-maker. A qualified member must review every check, verify the citations to the record, adjust ratings where the evidence is weak or incorrect, and formally sign off before this review is relied on for consistency, quality assurance, induction or training. Not legal advice.
🔧
Pre-filled Expert System
Click Open Pre-filled Expert System below to download a complete copy of the Expert System with every answer, evidence note, and score already filled in from this analysis. Open that file in your browser to review each answer, adjust where the AI got it wrong, add your own notes, and export the final signed-off PDF report.
·Detailed Findings by Practice
All 32 assurance checks · AI-generated evidence mapping · Confidence indicated per finding