Understanding the AI usage report
Our new AI usage report makes it easy to see how your org is using Writer and how you're tracking against your plan's usage limits.
In this article:
Accessing the AI usage report
To access the AI usage report, go to Admin > Reporting > AI Usage.
The AI usage report tracks 3 different metrics over time:
- Words generated: any content generated by Ask Writer, all chat apps or text generation apps (custom or prebuilt), highlights, commands, or the keep writing feature
- Recaps transcribed: any video or audio files which are run through our third-party transcription services (excludes already-transcribed videos on YouTube, Vimeo, etc, as well as txt/srt files uploaded by users)
- Words rewritten: any content generated via Writer-built rewrite, voice rewrites, and custom rewrites
Access roles & reporting
Your Writer access role will determine what kind of data you can view within the AI usage report.
Org admins have full access to all reporting data and can filter or group by any criteria.
Team admins have access to all reports, but they may only view data for the teams they are admins for, and the users within those teams. For example, a team admin would be able to group by "Team" but the only teams they would see in the report would be those for which they have team admin status. For all views, team admins will only see users on the teams for which they're team admins.
Understanding the AI usage report
Grouping your report
The date range (configured in the top left corner) determines which data sets are included in the report. The first Group by menu determines how you'd like that data "chunked" within your report. For example, if you're reporting on the past 2 months of activity, you may wish to view that data on a week over week basis. If you're reporting on the past year of activity, it may make more sense to view that data on a month over month basis.
The second Group by menu allows you to further group your data by various categories:
- By team: a stacked bar chart, with each "pill" representing a team.
- By billing group: stacked bar chart, with each "pill" representing a billing group.
- By feature: stacked bar chart showing where the usage occurred. Since some features generate words, while others rewrite words, the metrics may differ based on feature.
- By surface: stacked bar chart showing where in the Writer platform the usage occured: the Writer desktop app, an extension, etc.
- By user: a stacked bar chart, with each "pill" representing a user.
If you group your report by Time period, the data will be presented as a line graph representing activity over time.
If you group your report by team, billing group, feature, surface, or user, the data will be presented in a stacked bar chart, with a color-coded stacked bar for each of the three metrics.
When viewing a stacked bar chart, only the top ten usage sources will be included with the bar chart, with the rest of the data viewable in the table beneath the graph.
In the screenshot above, one month of data is grouped by week. The data is then grouped by feature, and hovering over an item in the stacked bar chart displays the feature represented. For the week of July 27th, we've hovered over the pink stacked bar, which represents Words Generated according to the rubric in the bottom left corner. Since we've grouped this data by feature, hovering over a pill of the pink stacked bar shows us that 269 words were generated by the image analyzer feature that week.
Filtering your report
In addition to grouping the data in your chat by certain categories, you can also apply filters to your chart to limit the data which is displayed in your report:
- Filter by team
- Filter by billing group
What's the difference between filtering and grouping?
Grouping data does not change anything about what data is included in your report – only the way in which it is displayed. Applying a filter to your report means your report only includes the data for that particular filter trait. In other words, grouping a report by team will display the words generated by your product team, your marketing team, and anyone else, allowing you to compare activity across teams. However, filtering a report by team will only show you the words generated by your product team, allowing you to track just the product team's activity over time.
Exporting your report
In the top right corner, select the ••• menu to export your report as a CSV or XLSX file.
Frequently asked questions
Who can access the AI usage report?
Org admins have full access to the AI usage report and can group or filter the data however they want.
Team admins have access to the AI usage report but can only view data from the teams they belong to.
How many years of data can be reported on?
The AI usage report can include data up to 2 years in the past.
Why can't I see more than ten pills in a stacked bar chart?
Stacked bar charts are organized from high to low usage. Only the top ten usage sources will be included in the bar chart, with the rest of the data viewable in the table beneath the graph.
Do these reports cover AI Studio or API usage?
API usage is not yet included in this report. To view custom app usage (including custom apps built by Writer and those built by users in AI Studio), group or filter your reports by feature.
What's the difference between filtering and grouping?
Grouping data does not change anything about what data is included in your report – only the way in which it is displayed. Applying a filter to your report means your report only includes the data for that particular filter trait. In other words, grouping a report by team will display the words generated by your product team, your marketing team, and anyone else, allowing you to compare activity across teams. However, filtering a report by team will only show you the words generated by your product team, allowing you to track just the product team's activity over time.
What are the two columns that appear when I group by feature?
Some Writer features generate words, while others rewrite words. To offer extra clarity, grouping by feature will create 2 columns to distinguish between the two data sets.
Words generated can include prebuilt applications (prebuilt), custom applications (built by Writer or by users in AI Studio), and other features (highlights, commands, keep writing).
Words rewritten can include voice rewrite (those generated by voice profiles), Writer rewrite (prebuilt styles like Simplify or Polish), or custom rewrite (rewrite options created by Writer, trained on your organization's writing).
When I group by feature, does the "custom application" category refer to apps built by Writer or apps built by customers in AI Studio?
Custom application data includes both data from Writer-built custom apps and customer-built apps.