The savings ledger: verified vs projected
What the two states mean, why the split matters, and how a saving becomes verified.
Updated Aug 3, 2026
The savings ledger (Analytics → Savings ledger) is the single source of truth for every pound the product claims. Each row is one saving event with its evidence and a status.
Projected
A projected saving is booked automatically when an agent completes a real run. It's modelled from your savings model (rate × minutes × realisation factor). It's a reasonable estimate — but it's still an estimate, and it's clearly labelled as such.
Verified
A verified saving is one a human has confirmed against the evidence. Toggling a row to verified is an audited action, and it's the number you can put in a board pack. The analytics header always shows the verified-vs-projected split so nobody confuses the two.
Why the split exists
Most "AI ROI" numbers are projections dressed up as fact — the gap between expected and realised return is exactly the credibility problem ProcessTwin exists to solve. Keeping projected and verified visibly separate is the honest way to report, and steering a customer toward their first verified saving on real data is what turns a pilot into a paying account.
Write-back
A verified saving can be posted to Xero or QuickBooks as a draft journal, so the ROI shows up in the accounts where finance already looks (live with finance keys connected; a preview otherwise).