Glossary
Learned rule
A rule an agent proposes from a pattern ('Great Lakes bills pre-rolls per 5-pack') that becomes active only after approval, is versioned, and lives in a governance tab — not in a model's memory.
Why it matters for reconciliation: the same quirk, such as a vendor that bills pre-rolls per 5-pack, will recur every month. Recording it as a rule saves a person from fixing the same mismatch repeatedly, while requiring approval keeps a wrong pattern from becoming a standing behavior.
Example: an agent notices that one vendor's invoices count pre-rolls per pack while the manifests count singles. It proposes a rule, a person approves it, and the rule is versioned. Later invoices from that vendor are mapped correctly and the rule appears in the governance tab where it can be changed.
Where this comes up on budally.com
- A connector is a checkbox. So is a model. Here's what the moat in cannabis retail ops actually is.
- Agents draft, humans approve: what an AI audit trail should actually contain
- Learned rules you can read: how 'tell it once and it remembers' should actually work
- Michigan's 24% wholesale marijuana tax: book it in cost, never on the receipt
Related terms
- Agent (BudAlly)
A scheduled or triggered process that reads ledger data, applies rule packs and drafts proposals — invoices, reorders, prices, vendor credits, campaigns — into the approval queue. Agents never execute; a named person approves each proposal.
- Mapping review
The screen where state-system items, POS SKUs and GL accounts are linked, with a confidence score; links at or above 0.90 are auto-accepted, the rest reviewed by a person.