Skip to content
CoefficientAnalytics

AI Analytics

AI Analytics best practices

A few habits make AI Analytics noticeably more useful day to day. This article covers the most common ones.

Know what time window you're looking at

AI Analytics answers over a recent rolling window of your data rather than a range you type into the question. Check the time window shown alongside an answer before comparing it to a specific period you have in mind, so you're not comparing numbers across different windows without realizing it.

Use follow-up questions instead of starting over

If an answer prompts a "why," ask that directly as a follow-up rather than framing an entirely new question from scratch. Follow-ups retain the context of the previous question, which usually produces a more directly useful answer.

Check the sourcing before acting on a surprising answer

If an answer is unexpected — a number much higher or lower than you'd guess — check the underlying metric and time window shown alongside it before assuming something is wrong operationally. Occasionally a surprising answer reflects a misunderstood question rather than a real change in the data.

Don't ask questions outside your connected data

Questions about data from a source you haven't connected to the workspace won't have an answer to draw from. If you're missing a data source relevant to a question you want to ask regularly, connecting it is usually worthwhile.

Share useful questions with your team

If a question turns out to be one you ask repeatedly, it's often worth turning into a permanent dashboard widget as well — AI Analytics is best suited for ad hoc and exploratory questions, while a recurring question you check daily may be better served by a stable dashboard view.

We use cookies to run this site and, with your consent, to understand usage and improve our marketing. See our Cookie Policy.