Skip to content

Turn data into something you can use.

Create data and machine-learning projects with executable work, evaluation results and downloadable artifacts.

Start free Explore pricing

From the brief to the result.

01

Describe the question

Set out the data, target and result you need. Make the intended evaluation clear.

02

Build and evaluate

Generate the project, run supported training or prediction work and inspect the separate evaluation results.

03

Keep the evidence

Review reports and output artifacts, then refine the project or take the files into your own workflow.

Results depend on the data and evaluation setup. Quant trading strategy builds are planned; they are not an available output type today.

From your data to an answer you can inspect.

Define the question, test the approach and keep the evidence behind the result.

Frame the question

Give the data a clear purpose.

Tell the AI what each row represents, which question matters and what a useful result would look like. The project can explore patterns, classify records, predict a value, forecast a series or discover groups. Its report makes the dataset and method visible.

  • Start with the data available to the project and a defined target or analysis question.
  • Inspect the data profile: columns, missing values and relevant distributions.
  • Set the task and evaluation metric before judging the result.
Example energy dataset with a data table, demand series and temperature distribution

Test the result

Compare models against a baseline.

For supported supervised tasks, the platform separates holdout data before the builder sees the training data. It runs the generated prediction code, recomputes the metric and compares baseline models on the same split. Review that independent result alongside the training estimate and verification findings.

  • Separate the training estimate from the platform-recomputed holdout result.
  • Compare against baseline models and inspect checks for data leakage.
  • Use the findings to refine the approach; future accuracy is not guaranteed.
Illustrative model comparison showing mean absolute error for a baseline, linear model and boosted trees

Keep the evidence

Read the report. Inspect the work.

The deliverable includes executable work and a report explaining the problem, data, method and results. Inspect the charts, error analysis and limitations, then use the available project downloads to continue in your own workflow or request a revision in the build conversation.

  • Review findings and limitations beside the supporting charts and evaluation.
  • Inspect training and prediction code with the available output artifacts.
  • Use the reproduction instructions to understand how the result was produced.
Example energy-demand report with observed and forecast values, findings, limitations and supporting files

The workflow

Keep the work connected.

  1. Describe the result

    Explain the audience, desired output and constraints. Refine the direction before moving forward.

  2. Review the work

    Inspect the generated files, available preview and the checks relevant to the output.

  3. Make the next version

    Ask for changes, then export or use the delivery workflow supported by that project.

Explore what’s next.

See all builders

AI Website Builder

Create landing pages, storefronts and content sites. Describe the audience and the experience, then review and refine the build.

Explore AI Website Builder

AI App Builder

Build web, mobile and desktop apps around the job they need to do. Preview the experience, refine it and prepare the appropriate package.

Explore AI App Builder

AI Game Builder

Turn a game idea into a playable project. Shape the rules, characters and world, then test and iterate on the experience.

Explore AI Game Builder
Questions & answers

A little clarity before you start.

What kinds of data projects can I build?+
The data workflow supports analysis, classification, regression, forecasting and clustering. Describe the question, available data and the result you want to evaluate.
What should I include in my brief?+
Explain what each row represents, which field you want to predict or analyze, and how you will judge a useful result. Include relevant constraints and the data available to the project.
How are model results checked?+
For supported supervised tasks, the platform separates holdout data before the builder sees it and recomputes evaluation results. It also checks against baselines and looks for data leakage. Inspect the findings alongside the report.
What do I receive?+
A project can include training and prediction code, a report, evaluation results and output artifacts. Review the generated files and use the available download options to continue working with them.
Can I improve an existing result?+
Yes. Use the build conversation to refine the target, explain an error or request a different approach. Review the new evaluation results to see whether the change helped.
Does a good evaluation guarantee future accuracy?+
No. Evaluation describes performance on the data and test setup used. New data, changing conditions and differences from the intended use can change the result.

What will you make?

Create a build, or put a specialist team to work on your next task.

Describe your idea Explore agent teams