Rule Mining

The problem

Most business systems have a layer of authored rules that handle the long tail of decisions the application cannot decide for itself. "If vendor X and category Y, route to approver Z." "If amount > €5,000 and project = construction, require CFO sign-off." Authored rules are auditable and well-understood, but they have a discoverability problem: someone has to know that the rule should exist before they can write it.

The patterns the rules should encode are already in the data. Vendor X plus category Y has routed to approver Z 47 times out of the last 50 occurrences — that is a statistically significant pattern with 94% support. Nobody has written it as a rule yet, but the system could surface it as a candidate. Once surfaced, the team can promote it to a rule or dismiss it, and the decision is recorded.

Rule mining solves the discoverability problem. The system finds the patterns hiding in the operational data and presents them for human review.

How it works

Aito's _relate operator returns conditional probabilities between attributes — and crucially, ranks them by lift. Lift is the ratio of conditional probability to base probability; a lift of 10× means the outcome is ten times more likely when the pattern holds than on average. Lift alone is not enough: a high lift over three invoices is still thin evidence, so each candidate also carries its support, how many invoices it covers and how many of those it gets right.

The accounting demo's rule-mining view runs _relate with $patterns once per frequent GL code and approver, over the fields known when an invoice arrives, then counts each candidate's support exactly with _search. Each candidate becomes a row with the conditional pattern, its lift and its support. In both the accounting and ERP demos, Promote requires explicit sign-off and writes an audit-trail entry, and the ERP demo's Dismiss records a rejection; nothing mined is policy until someone promotes it.

{
  "from": {
    "from": "invoices",
    "where": {
      "customer_id": "CUST-0000"
    }
  },
  "where": {
    "gl_code": "1600"
  },
  "relate": {
    "$patterns": {
      "$related": {
        "relate": [
          "vendor",
          "category",
          "vendor_country",
          "amount_band"
        ],
        "k": 8,
        "to": {
          "gl_code": "1600"
        }
      }
    }
  },
  "select": [
    "related",
    "condition",
    "lift"
  ],
  "orderBy": "lift",
  "limit": 6
}

For the full architecture, see the technology overview. For the broader narrative across multiple use cases, read The Predictive Application.

See it live

This use case runs in 2 live demos today, the 📋 ERP and 🧾 Accounting demos. Each link opens the view that shows it, where you can inspect the queries that produce the result.

• ERP demo →
• Accounting demo →

Frequently asked