Scenarios
See what a change would have saved on last month's AI calls.
Every figure, and how it is computed
- Would have saved
- What the AI calls the change touches cost, minus what they would cost after it, a month.
- Cost per outcome
- The whole workflow's cost per outcome, before and after the change.
- Quality
- The share of results scoring 0.8 or better. After the change: not known until it has run.
- AI calls affected
- AI calls the change touches, and their share of all spend.
- With and without the change
- Daily spend on the AI calls the change touches, as it was and as it would have been.
- The price ladder
- For a routing opportunity: the same exact tokens on every model the rate card prices, cheapest first.
- Why this works, and what could break it
- The opportunity's own reasons, beside the change.
What you do here
- Move the slider; the figures recompute from the records.
- Approve, assign an owner or reject from here.