_relate
Conjunction rules from _relate with $patterns. Aito mines the AND-rules a human would write for each output an AP clerk codes — GL code and approver — from intake inputs only (vendor, category, amount band). No output is used to predict another, so the rules actually fire at routing time. Support ratios are exact historical counts, not ML estimates.
Example query
{
"from": {
"from": "invoices",
"where": {
"customer_id": "…"
}
},
"where": {
"gl_code": "1600"
},
"relate": {
"$patterns": {
"$related": {
"relate": [
"vendor",
"category",
"vendor_country",
"amount_band"
],
"k": 8,
"to": {
"gl_code": "1600"
}
}
}
}
}
Verify yourself
Every claim on this page traces back to an Aito query. No separate model file — Aito predicts directly from its index.
Note: GL codes (4100, 5300, 6200…) are illustrative — real Finnish Liikekirjuri uses different numbering. Chosen for demo readability.