Connect BigQuery

Connect BigQuery to analyze tables that live only in your own database, such as orders, members, and logs, alongside your ad performance. You can ask questions right in chat that ad platform data alone can't answer, like ROAS based on actual revenue, new vs. repeat customers, or performance by membership tier.

What you'll need

• A Google account with access to your BigQuery data — you'll sign in with this account when connecting.

• Read access to the datasets you want to query — ask your Google Cloud admin for permission to view data (BigQuery Data Viewer, roles/bigquery.dataViewer).

• Job User access in the project that will run queries — ask your Google Cloud admin for permission to run queries (BigQuery Job User, roles/bigquery.jobUser). Without it, you can see the data but queries won't run.

• The BigQuery API enabled in that project — if anyone has used BigQuery in the console for that project, it's already on.

Step 1 — Connect your Google account

1. On your workspace's Data sources page, click Connect on the BigQuery card.

2. Click Sign in with Google and sign in with the account that can access your data.

3. On the consent screen, be sure to check the BigQuery permission. If you skip it, the connection won't go through. Sign in again and check that item.

Step 2 — Choose a project

Choose the project your data is in. The project you choose here is where queries run and where scan costs are billed.

Step 3 — Choose datasets and set scan limits

1. Check the datasets you want to allow queries on. Nothing is selected by default — only the datasets you choose are opened.

2. Switch projects at the top to include datasets from multiple projects (you need query access in each project).

3. Scan limit per query (GiB) — default 10. Queries that would exceed this are blocked before they run. Raise it if your tables are large.

4. Daily scan limit (GiB) — default 100. This is the total your workspace can scan per day, and it resets every day at 9 AM Korea time. A per-query limit alone can't stop many small queries from adding up, so the two work together.

5. Notes about your data (optional) — describe your frequently used tables and what the code values mean (up to 2,000 characters). The AI can't tell what status = 3 means from column names alone. Example: orders = order ledger, status 3=paid / users.grade = membership tier (1–5)

6. Click Connect with N dataset(s) and you're done.

Change settings after connecting

Click the Settings (gear) icon on the BigQuery card among your connected services to change these without reconnecting:

• Add or remove datasets the AI can query

• Per-query and daily scan limits

• Notes about your data

The daily limit must be equal to or greater than the per-query limit to save. If the daily budget were smaller than a single query, every query on that connection would be rejected from the start.

Example requests

Once connected, just ask in plain language in chat. The AI looks up the allowed datasets and table structures on its own, so you don't need to memorize table paths.

• "Show me the datasets and tables I can access right now."

• "Show me what fields are in the orders table."

• "Combine last month's orders with ad spend by platform and calculate ROAS based on actual revenue."

• "Split new and repeat customers and compare each platform's contribution."

• "Show purchase conversion rates by membership tier for each campaign."

Troubleshooting

It says I need to check the BigQuery permission

You moved past the Google consent screen without checking BigQuery. If you skip it, the connection won't go through. Sign in again and check that item.

It says no BigQuery projects are available

Either the Google account you signed in with can't access the project that uses BigQuery, or the BigQuery API isn't enabled in that project. The most likely cause is the wrong account (for example, a personal account instead of your work account), so check that first. If the list is still empty, ask your Google Cloud admin for project access and to enable the BigQuery API.

It says I don't have permission to run queries in this project

You have permission to view data, but not to run queries. Ask your Google Cloud admin for permission to run queries in that project (BigQuery Job User, roles/bigquery.jobUser), then try again. Data Viewer alone isn't enough to run queries.

I can't select datasets from different regions together

Data stored in different locations (regions) can't be analyzed together. Choose datasets from the same region only. If needed, ask your data team for help.

My query is blocked for exceeding the scan limit

Ask the AI to query again "with a shorter date range" or "with only the fields I need." If that's still not enough, raise the per-query limit in Settings.

I've hit the daily scan limit

It resets automatically every day at 9 AM Korea time. If you need to keep going before then, raise the daily limit in Settings.

The AI misreads what column values mean

Describe that table and its code values in Notes about your data in Settings. These notes are sent to the AI with every query, which noticeably improves accuracy. On the other hand, long notes add cost to every query, so keep them short and focused on the tables you use most.

The AI says it can't see a certain dataset

That dataset wasn't selected when you connected. Add it in Settings. The change takes effect immediately, with no need to reconnect.