Aito v2 Use Cases
Six things teams build on the v2 API, each written around queries you can run as-is against the live demo dataset (the same data as the sandbox). The GL coding page prints responses recorded from a test run of its queries on that data; on the other pages the printed responses show the shape — run the query for the current numbers.
Automated GL Coding
Predict general-ledger accounts and approvers for incoming invoices — calibrated, explainable AP automation
Support Ticket Triage
Classify messages, suggest answers from resolved history, and scope per tenant with population restriction
Multi-Label Auto-Tagging
Score every label independently with predict field.$feature — catalog and content enrichment
Market Basket Analysis
Mine co-occurring itemsets live with relate $patterns — association rules as a query
Full-Text Search
BM25-ranked $search with boolean syntax and highlighting — no separate search engine
Behavioral Analytics
Frequency vs lift per user or segment — what's bought, and what's characteristic
Knowledge Graphs
Node classification and link prediction over ordinary collections — $refs, predict over a link, and $why, with no graph database
Personalised Recommendations
Rank products by the probability of a purchase for this user and basket, with exclusions and $why
Smart Search
BM25 relevance for the matches, purchase probability for the order: search personalised per user
Personalised Autocomplete
Complete a typed prefix from the search log, re-ranked by what this user tends to search for
Predictive Cart Autofill
Predict a shopper's basket as a multi-label prediction, and measure it against popularity first
Product Analytics
Conversion, category comparison, buyer segments, substitutes and weekly demand for one product
Quality Monitoring
Measure a prediction with _evaluate: baselines, error cases, calibration, time splits and profile A/B tests
Price and Demand Analytics
Estimate price and units sold in context, trace a demand curve, and explain the estimate