Skip to main content

Connecting BigQuery to AlphaSense

Kristin Mize
Updated

The BigQuery connector lets you ask questions about your Google BigQuery data directly in AlphaSense. You connect your Google account, register the BigQuery project you want to query by project ID, and then ask questions in plain language — AlphaSense discovers the tables you have access to and runs the analytical queries for you, on demand. Your BigQuery data is queried live and is never copied into AlphaSense.

Contents

Before You Begin

To connect BigQuery you'll need:

  • An AlphaSense account with the BigQuery connector enabled for your organization. Please see the Admin setup section below for enablement instructions.
  • A Google Cloud project that holds the BigQuery data you want to query, with billing enabled. BigQuery requires a billing‑enabled project to run queries (a free monthly usage tier still applies); the credit‑card‑free BigQuery sandbox works only within sandbox limits (Google: BigQuery sandbox & billing).
  • A Google account with access to that project's data. You'll connect with this account, and AlphaSense will only ever see what this account is already permitted to see.
  • The required Google Cloud APIs enabled on the project (see admin steps below).
  • If your Google account is managed by Google Workspace: your Workspace admin must allow AlphaSense to access Google data — some organizations block third‑party apps by default (Google: control which apps access Workspace data).

You only ever see data your own Google account can access. AlphaSense acts on your behalf and cannot reach anything you couldn't reach yourself.

How to Connect BigQuery

Connecting has two parts: a one‑time setup by a Google Cloud / Workspace admin, then a per‑user connection in AlphaSense.

Step 1 — Admin setup (one‑time, per project)

These are admin‑only actions — they require Google Cloud IAM access to the project. If you're an end user without that access, share this section with your Google Cloud / Workspace admin to complete before you connect.

Prerequisites

  1. A GCP project that owns the BigQuery data you want to expose. You'll grant your end users access to query this project.
  2. Project-level admin access (Owner / IAM Admin) on that project — only for the one-time API-enable and role-grant steps.
  3. End users' Google accounts must be in your Google Workspace and able to sign into Google Cloud Console.

A person with admin rights on the Google Cloud project must:

1. Enable the required APIs on the project:

  • BigQuery API (and BigQuery Storage API) — usually already in place for existing BigQuery projects, but worth double checking.
  • Conversational Analytics API (recommended) — powers the natural‑language questions (Google: enable the Conversational Analytics API).

The Conversational Analytics API is the preferred path; it gives users richer answers (analytical narrative, charts in a future release). If you choose not to enable it, AlphaSense automatically falls back to direct BigQuery execution — users would still get results, just with shorter answers
and no AI narrative.

Cost note. Enabling these APIs is free. You only pay for the queries AlphaSense runs on your behalf:

  • BigQuery: standard query costs, billed to this project.
  • Conversational Analytics: token-based pricing (Gemini), also billed
    to this project.

Set up billing alerts on the project if you want a hard ceiling.

2. Grant each connecting user access to the project. A user needs all three of the following, because running a query, reading the rows, and using natural‑language analytics are separate permissions (Google: BigQuery access control, Google: Conversational Analytics IAM):

Grant this role So the user can
roles/bigquery.user Run queries and discover the catalog (list datasets & tables)
roles/bigquery.dataViewer Read the actual table data (bigquery.user alone can't read rows)
roles/geminidataanalytics.dataAgentStatelessUser Use the natural‑language (Conversational Analytics) path

In gcloud:

# For an individual user:
gcloud projects add-iam-policy-binding YOUR_PROJECT_ID \
  --member="user:alice@yourcompany.com" \
  --role="roles/bigquery.user"

gcloud projects add-iam-policy-binding YOUR_PROJECT_ID \
  --member="user:alice@yourcompany.com" \
  --role="roles/geminidataanalytics.dataAgentStatelessUser"

# For a Google Group (recommended for >5 users):
gcloud projects add-iam-policy-binding YOUR_PROJECT_ID \
  --member="group:bq-readers@yourcompany.com" \
  --role="roles/bigquery.user"

Or in the Console: IAM & Admin → IAM → Grant Access → add the principal, add both roles, save.

If you don't want users to see every dataset in the project, use roles/bigquery.user for project-level listing/jobs and then a narrower roles/bigquery.dataViewer only on the specific datasets you want exposed.

AlphaSense respects dataset-level row/column security; users can only see data their token authorizes.

3. (Workspace‑managed orgs only) In the Google Admin console → Security → Access and data control → API controls, mark AlphaSense as allowed so users aren't blocked from connecting (Google: manage third‑party app access).

Step 2 — Connect your Google account (for each user)

  1. In AlphaSense, open the Integration Center and, under Structured Data Tools, select BigQuery.
  1. Click Connect BigQuery and sign in with Google. Google shows a consent screen — "alpha‑sense.com wants access to your Google Account" — requesting to see, edit, configure, and delete your Google Cloud data and see your account email (Google's standard wording for this level of access). AlphaSense uses this only to discover and query your BigQuery data on your behalf. Click Allow.
  1. Your Google account is now linked to AlphaSense.

Step 3 — Register your BigQuery project

Linking your Google account is not enough on its own — you then tell AlphaSense which BigQuery project to query by registering it. On the BigQuery page, click Add Project to open the Connect a BigQuery project dialog:

  1. (Optional) Give the project a friendly Name.
  2. Enter the GCP Project ID that hosts your datasets — "the Google Cloud project ID where BigQuery is enabled."
  3. Click Connect. AlphaSense validates access at registration time: it confirms your connected account can reach the project's datasets and that the natural‑language (Conversational Analytics) capability is available for that project. If access is missing or the API isn't enabled, registration fails with a clear message — see Troubleshooting. Then click Done.

What's Accessible

The BigQuery connector can reach the datasets and tables in the Google Cloud project you have registered and that your connected Google account is permissioned for — nothing more. It discovers the catalog and runs read‑only analytical queries against it in real time. Your registered BigQuery data is available to AlphaSense's Generative Search and Smart Summaries.

Capabilities

Capability What it does
Catalog discovery Lists the datasets and tables available to you in the project
Natural‑language querying Answers questions about your BigQuery data by generating and running SQL for you, via Google's Conversational Analytics (overview)

AlphaSense performs read/query operations only — it does not create, modify, or delete your BigQuery data.

How It Works

  • You register the project. You explicitly register the BigQuery project by its project ID, and AlphaSense validates your access to it — and that the analytics capability is enabled — before it can be queried. Nothing is reachable until you've both connected your account and registered the project.
  • Real‑time, not ingested. AlphaSense queries BigQuery live when you ask a question. Your data is not copied into or indexed by AlphaSense's search index.
  • On your behalf. Every query runs using your connected Google identity, so results are automatically limited to what your account can access, and permissions are checked live at query time by Google.
  • Broad grant, narrow use. The connection uses a broad Google Cloud access grant (required by the Conversational Analytics capability), but AlphaSense uses it only to discover and query your BigQuery data.
  • Short‑lived retention. Query results are retained only in AlphaSense's observability traces for 10 days, then automatically deleted.

Items to Note

  • Users may only query the registered projects. You can query a project once you've registered it, and each user may only query 1 project at a time.
  • Access is the #1 blocker. If a question returns "BigQuery analytics isn't available for this project" or no data, it almost always means the connected account is missing one of the three grants above, an API isn't enabled, or the project has no data — not a connection failure.
  • Google BigQuery limits apply. Very large result sets or unusually long‑running queries are subject to Google's standard BigQuery limits (Google: BigQuery quotas & limits).
  • Read‑only. The connector never writes to your data.

Modify, Revoke, or Disconnect

  • In AlphaSense: open the BigQuery page in the Integration Center (it shows your Account, Status: Connected, and Connection Type: Individual), click Disconnect, and confirm — "Disconnecting will unlink your BigQuery from AlphaSense." This revokes access and stops all live querying. Because nothing is ingested, there's no stored content to remove; any query results held in observability traces are automatically deleted after 10 days.

Troubleshooting & FAQs

"Google hasn't verified this app" when I try to connect.

This appears during the pre‑verification window for the requested access level. If you're an approved tester, continue via Advanced → Go to AlphaSense; otherwise wait until verification is complete (Google).

"Access blocked" / I can't reach the Google sign‑in.

Your Google Workspace admin may be blocking third‑party apps. Ask them to allow AlphaSense in the Admin console API controls (Google).

AlphaSense says analytics "isn't available for this project," or returns no data.

Your connected Google account likely lacks one of the required grants on the project — you need to be able to run queries, read table data, and use Conversational Analytics (roles/bigquery.user + roles/bigquery.dataViewer + roles/geminidataanalytics.dataAgentStatelessUser). Also confirm the required APIs are enabled and the project actually contains data. Ask your project admin to complete Step 1.

Registering a project fails ("project not found," "access denied," or "enable the API").

Registration checks your access before saving the project. A failure means the project ID is wrong, your connected Google account can't reach that project's data, or the required APIs (BigQuery + Conversational Analytics) aren't enabled on it. Fix the access/APIs (see Step 1) and register again.

AlphaSense doesn't know which project to use.

Reference the registered project you mean in your question.

Can AlphaSense change or delete my BigQuery data?

No. It only reads and queries.

Security

AlphaSense queries your BigQuery data on your behalf, honoring your existing Google permissions, and does not ingest it. For details on AlphaSense's security posture and certifications, see trust.alpha-sense.com.

Sources

Was this article helpful?

0 out of 0 found this helpful
Return to top

Related articles