Dravity logo

Auto ML Agent

Forecast sales, anticipate churn and uncover opportunities with an AI agent that helps you analyse data, train models and put predictions to work. All inside your Snowflake account.

Turn your data into your next decision.

Bring predictive analytics into everyday decisions. Forecast revenue and demand, predict customer churn, and uncover patterns with a conversational Auto ML agent. No code or dedicated data science team required.

Guided machine learning Compare models Predict with confidence
Dravity Auto ML Agent preview
BUILT FOR BUSINESS QUESTIONS

Move beyond reporting. Start anticipating.

Use AI-powered predictive analytics to answer important questions across sales, finance, operations and customer teams, so you can plan earlier and act with greater clarity.

01

Customer churn

Identify customers at risk of churn so your teams can focus retention efforts before they leave.

02

Sales and demand

Forecast sales, revenue, inventory needs and demand across stores or product lines to support better planning.

03

Customer segmentation

Group customers by patterns in their data to support more relevant business decisions.

04

Business outcomes

Estimate prices, delivery times, customer value and other measurable results.

A GUIDED PATH TO PREDICTIONS

From dataset to prediction, one guided journey.

Build predictive models without writing code. Describe your goal to the Auto ML Agent or follow the five-step Model Wizard. Both use the same machine learning engine.

01

Choose your data

Start from an existing Snowflake table or upload a CSV dataset.

02

Let AI analyse it

Review the proposed task, target and features, with data quality checks before training.

03

Train and compare

Train a baseline or compare candidate machine learning models on an AutoML leaderboard.

04

Review and save

Inspect model results and choose the candidate you want to keep.

05

Predict and forecast

Score rows or entire tables, create forecasts and use the results in dashboards.

AUTO ML AGENT Illustrative conversation
Forecast daily sales per store for the next 60 days.
Proposed analysis Sales forecasting

Review the dataset, date column, store grouping and training plan before you approve compute.

Data checks Compare models Review results
YOUR DATA SCIENTIST, ON DEMAND

Just describe what you want to predict.

The conversational Auto ML Agent helps select a prediction task, check data quality, train and compare models, and explain the results in business language. You approve the plan before training begins.

  • Ask naturally, without writing Python or building ML pipelines.
  • See progress while training continues in the background.
  • Return to your saved models to predict, forecast or retrain.
FOUR WAYS TO LEARN FROM DATA

The right prediction for the right decision.

Explore AI-powered classification, regression, clustering and time series forecasting in Snowflake, from predicting business outcomes to forecasting demand across many locations.

Classification

Predict a category or outcome, such as customer churn, fraud or lead conversion.

Regression

Estimate numerical results like price, delivery time, demand or customer value.

Clustering

Discover groups and patterns across customers or products.

Forecasting

Forecast sales, revenue, traffic and stock for one time series or hundreds of series together.

Forecast at business scale.

Model daily, weekly or monthly data, compare forecasting approaches, and handle multiple stores or product lines with panel forecasting when the data supports it.

BUILT FOR REAL-WORLD DECISIONS

Powerful predictions. Practical control.

Bring machine learning into business workflows with Snowflake-native access controls, data-quality checks, approval before training and reusable predictions.

Checks before training

A deterministic data-quality gate prevents training on datasets that fail required checks.

You approve compute

The agent asks for confirmation before training spends warehouse resources.

Runs in your Snowflake account

Training happens on your warehouse, with role-based access and per-user ownership.

Protected and reusable models

Saved models are encrypted and signed. Retrain on new data while retaining earlier versions.

FROM PREDICTION TO ACTION

Results your teams can actually use.

Save predictions and forecasts as Snowflake tables, then connect those results to dashboards, reports and everyday analytical workflows.

PERIODSTOREFORECAST
Week 01NorthForecast value
Week 02NorthForecast value
Week 03NorthForecast value

Illustrative format only, not actual model output.

FREQUENT QUESTIONS

Predict with clarity.

Answers about no-code machine learning, sales forecasting, model security and using predictions inside your Snowflake dashboards.

Do I need to know Python or machine learning?

No. The Auto ML Agent works through conversation, and the five-step Model Wizard offers a visual alternative. You can still review model choices and results.

Can Dravity forecast multiple stores or products?

Yes. Dravity supports forecasting a single series and panel forecasting for many series, subject to the structure and quality of the selected data.

Where does model training happen?

Training runs on your Snowflake warehouse, not in the app container. Compute-intensive training requires your confirmation.

Can I use predictions inside dashboards?

Yes. Predictions and forecasts can be written to ordinary Snowflake tables for use in Dravity dashboards or other Snowflake-based workflows.

How are datasets and saved models protected?

The platform enforces user permissions and ownership. Saved model artifacts are encrypted and signed, and failed data quality checks can block training.

FROM DATA TO THE NEXT DECISION

Turn predictions into better business decisions.

Explore Dravity Auto ML for customer churn prediction, sales forecasting and scalable predictive analytics, with model training inside your Snowflake account.