Examples

The MOSTLY AI Assistant empowers everyone to work with data like an expert. Connect to resources, create new data, build visualizations, and share artifacts using natural language.

Get data insights

The MOSTLY AI Assistant makes it possible for users to capture insights and understand their data as if they are working alongside a dedicated analyst. The Assistant can run code to transform the data, create visualizations, or offer suggestions about impactful analyses. MOSTLY AI empowers everyone to extract the most value from every dataset.

Example prompts

• Visualize the Female vs Male distribution in this dataset.
• Show me the correlation between the Age and the Income in this dataset.
• What type of analysis should I perform on this dataset to communicate my point?

Generate mock data

Users can prompt the Assistant to generate mock data for any use case. Describe your use case in conversation with the Assistant and give pointers to customize it exactly to your needs. Generate realistic test data and experiemnt with data model structures entirely using natural language.

Example prompts

• Create a dataset with 100 rows and 5 columns for first name, last name, nationality, company position, local office.
• Generate an HR dataset with 1000 rows and columns for employee ID, position, salary.
• Generate a list of flights departing from Vienna on a Sunday including their destination, departure time, and airline.

Enrich data and create artifacts

Enriching existing datasets is ideal if you want to use real-world knowledge used to train LLMs. Specify the dataset to enrich and how. Create a visualization and share it with other MOSTLY AI users or members of your organization.

Example prompts

• Add a country code column to the dataset.
• Enrich the dataset with the GDP per capita of the country.
• Add the population of the country to the dataset.

Access privacy-safe data

MOSTLY AI is the pioneer in privacy-preserving synthetic data generation. You can use the Assistant to create resources like connectors and generators, and synthesize a privacy-safe dataset that is can act as a drop-in replacement of the original when analyzed.

Example prompts

• Generate synthetic data with the US Census generator.
• Probe the US Census generator for 100 synthetic samples of only females.
• Generate a rebalanced US Census dataset with equal split of Male and Female.