Predict Any Field — With One Query

Aito's _predict endpoint returns the most likely value for any field in your dataset, along with a calibrated confidence score. No model training, no feature engineering, no deployment pipeline. Upload your data and start predicting.

Explainable predictions — Aito shows why it predicted David Green with 99% confidence, breaking down base probability and pattern match

Try Live Predictions

Run real prediction queries against our demo database. Every result includes a probability score — Aito tells you exactly how confident it is.

{
  "from": "invoices",
  "where": {
    "ProductName": "Cloud Services (AWS/GCP)",
    "TotalAmount": 5000,
    "InvoiceType": "Service"
  },
  "predict": "GLCode"
}

How It Works

1. Upload Your Data

Send your dataset to Aito as a table — CSV, JSON, or via API. No preprocessing required.

2. Write a Predict Query

Specify which field to predict and what you know:

{
  "from": "invoices",
  "where": {
    "ProductName": "Cloud Services (AWS/GCP)",
    "TotalAmount": 5000.00
  },
  "predict": "GLCode"
}

3. Get Results with Confidence

Every prediction includes a calibrated probability. Use it to set automation thresholds — route high-confidence predictions to automation, low-confidence to human review.

Key capabilities:

• Any field as target — predict GL codes, categories, assignees, or any other field
• Confidence scoring — calibrated probabilities on every prediction
• Multiple inputs — use any combination of known fields as context
• No retraining — add new data and predictions update immediately

Where This Is Used

• Accounting & Finance — GL code prediction, invoice routing, payment classification
• Process Automation — document classification, approval workflows
• Customer Support — ticket routing, priority assignment

Get Started

Start for free → Upload your data and make your first prediction in minutes.

Read the API docs → Full reference for the predict endpoint, confidence scoring, and evaluation.