It starts with a Spark

We review all customers who have signed up and will select beta applicants from a diverse range of industries and sizes. On July 17th, we expect to start releasing Spark to beta participants depending on if we are comfortable with the performance and quality of our results.

We have applied enterprise privacy and security practices to our development of Spark, including:

• Data is not used to train or improve third party generative AI models
• Data is not saved to disk or retained by third party generative AI models

You can find additional information about Mixpanel’s privacy and security practices here.

Spark is powered by OpenAI.

You can give Spark almost any question or request that involves an objective calculation or modification of your data, for example, “how many videos were watched in the last month?” or “break down this data by country”.

You will not be able to ask subjective questions to Spark, which we like to call “why” questions. An example would be “why are users purchasing less items this month?”

It is recommended that you try to remove duplicate event & property data and give this data simple, distinguishable names. This will make it easier for Spark to build the most accurate report based on your questions.