The AI-native strategy company · Starting in India
Glossary · strategy1 term
Definition
The evidence ledger is the record that links every number and claim in a strategy1 deliverable to its source, its date and the agent that produced it. Each entry also carries a confidence grade, and entries are chained, so any later change shows. A reader can click a figure and see where it came from.
Explained in full on Evidence and trust.
Context
AI systems can produce numbers that look right but have no source. A ledger makes every figure checkable, so a board can ask how a number was made and get an answer, and a wrong input can be traced to every conclusion that rests on it.
In practice
The ledger is built at the Model stage of the S1 Method, when the Strategy Twin is built and every fact is tagged with its source and confidence. The Narrative agent cannot add facts that are not in the ledger, and every Engine Sprint deliverable carries an evidence extract.
FAQ
No. The Narrative agent that drafts the memo cannot add facts that are not in the evidence ledger, and the Research agent stores every claim with its source. A figure without a ledger entry does not reach a deliverable.
Each evidence ledger entry is linked to the one before it, so if a source, date or grade is changed after the fact, the change shows. That makes the ledger an audit record as well as a list of sources.
The evidence ledger records where each input came from. Calculations run in strategy1's own deterministic code, so the language model explains the numbers but does not calculate them.
A 72-Hour Bootcamp takes one problem your team would scope in weeks and works it to a decision memo your leadership signs. Fixed scope, fixed fee.