Looker Studio

Connect AI Peekaboo to Looker Studio, see every field the connector exposes, and avoid the most common setup mistake with sentiment data.

5 minute read · Last reviewed: August 14, 2026

You want your AI Peekaboo data flowing into a Looker Studio dashboard alongside your other client reporting, instead of pulling numbers manually.

Setting it up

Go to Settings → Integrations.
Generate a Looker API key.
In Looker Studio, add the Peekaboo community connector.
Paste your API key into the connector to authenticate.

The connector is included on every current self-serve plan, Starter, Peek, and Grow, so there's no separate add-on to buy.

Choosing a data mode first

Before any field list matters, the connector asks you to pick a Data Mode, and that choice decides which fields exist in your data source:

Summary (recommended) returns one row per prompt per day, already averaged across the AI models that ran. Scores are stable and row counts stay small, which is what most client dashboards want.
Detailed (per AI model) returns one row per AI response, so you can split results by model and by the source domain that was cited. Row counts are much larger.

Two more options sit alongside it: a Brand dropdown, which defaults to your most active brand, and an Include competitor data checkbox, which adds competitor rows next to your brand rows so you can compare them with the Entity Name dimension. A date range is always required.

Fields in Summary mode

Fourteen fields, one row per prompt per day.

Field	Type	What it means
`Date`	Dimension	Analysis date, typed as a real date so Looker Studio time controls work on it
`Brand Name`	Dimension	The brand the row belongs to
`Tracked Prompt`	Dimension	The prompt text that was analyzed, truncated to 100 characters
`Prompt ID`	Dimension	Unique prompt identifier. Use `COUNT_DISTINCT` on this for an accurate prompt count, since the same prompt text repeats across dates
`Search Intent`	Dimension	Intent classification of the prompt, for example `COMMERCIAL` or `INFORMATIONAL`
`Prompt Category`	Dimension	Your strategy-based grouping for the prompt
`Sentiment`	Dimension	Majority sentiment across the models that ran: positive, negative, or neutral
`Entity Type`	Dimension	Whether the row is your `brand` or a `competitor`
`Entity Name`	Dimension	Name of that brand or competitor, the field you group by for side-by-side comparison
`AI Visibility Score`	Metric	0 to 100 visibility score, averaged across models. Defaults to `AVG`
`Total Citations`	Metric	Brand mentions counted across the models that ran
`AI Responses Analyzed`	Metric	How many AI model runs sit behind that prompt and date
`Models Used`	Metric	How many distinct AI models ran for that prompt. Defaults to `AVG`
`Average Position`	Metric	Average rank position in the answer, where 0 means not mentioned. Defaults to `AVG`

Fields in Detailed mode

Also fourteen fields, but one row per AI response. Prompt ID and Models Used disappear, and Source Domain plus AI Model take their place.

Field	Type	What it means
`Date`	Dimension	Snapshot date
`Brand Name`	Dimension	The brand the row belongs to
`Source Domain`	Dimension	The domain cited in the answer, for example `forbes.com`, or `(not set)` when the response cited nothing
`Tracked Prompt`	Dimension	The prompt text that was analyzed, truncated to 100 characters
`AI Model`	Dimension	Which model produced the response, for example `gpt-4o-mini` or `gemini-2.5-flash`
`Search Intent`	Dimension	Intent classification of the prompt
`Prompt Category`	Dimension	Your strategy-based grouping for the prompt
`Sentiment`	Dimension	Brand sentiment in that single AI response
`Entity Type`	Dimension	Whether the row is your `brand` or a `competitor`
`Entity Name`	Dimension	Name of that brand or competitor
`AI Visibility Score`	Metric	0 to 100 visibility score for that single response. Defaults to `AVG`
`Total Citations`	Metric	Brand mentions attributed to that source
`AI Responses Analyzed`	Metric	Count of analyzed responses
`Average Position`	Metric	Rank position in that response, where 0 means not mentioned. Defaults to `AVG`

Sentiment must be a dimension, not a measure

Sentiment in AI Peekaboo is categorical: positive, neutral, or negative. If you add it to a Looker Studio chart as a measure (a numeric aggregation), Looker Studio will try to average or sum it, and you'll get a meaningless number instead of a readable breakdown.