On using AI
AI, used lightly.
We build tools, not chatbots. AI earns its place on Revel Sphere when it helps community members do something together — draft an event, sort a roster, find the next step. The rest is restraint.
We reach for the lightest model that does the job.
Most of what our tools do runs on a small, efficient model. The heavier ones are held back for the work that genuinely needs them.
We reuse what has already been computed.
Every time a tool runs, it leans on the same shared instructions. Rather than process those from scratch each time, we keep the computed result and reuse it — the way you reprocess a material rather than make it new. Reused context means less work per request, run after run.
We cap what could run away.
Our tools watch their own usage. As demand climbs they step down a tier and pause the non-essential work, so a busy week stays measured and within bounds.
We keep receipts.
We track how much context gets reused and what that reuse saves, and we hold ourselves to it. We would rather show a real number than make a green claim we cannot stand behind.
Receipts over rhetoric.
We embrace what AI makes possible for community work — and we stay honest about its weight. Have a question about how the tools run? We will answer it plainly.