- explain report terms
- summarize key points
- suggest follow-up questions
- help users understand risk/timing/action
- classify feedback
AI EXPLANATION
AI Explanation Layer
AI should help users understand a report, not invent the chart.- recalculate the chart
- change verdicts
- invent new chart data
- provide guaranteed future claims
- replace human review for high-stakes decisions
Usage model
AI explanation needs boundaries before it needs more chat.
limited explanations
safe report context
no raw private tokens
model routing
monthly cost caps
upgrade path
Example AI helper
The helper answers from an approved report context. It explains terms and next actions, but it does not create a new chart, change the verdict, or make guaranteed claims.
What does major luck cycle mean?Why is this a timing risk?What should I do first?What needs human review?
Service deliverables
The output is a controlled AI layer, not an open-ended fortune chat.
AI behavior spec
prompt boundaries
safe context design
refusal rules
usage quota model
UI sample