How to get your Google Ads data ready for agentic AI

Sync Google Ads campaigns, ad groups, and performance reports with Airbyte, including custom GAQL queries, so AI agents can analyze paid search.

|
Nov 16, 2026

Summarize with AI:

Airbyte syncs Google Ads resources, performance reports, and custom GAQL queries into your warehouse, and the Airbyte context layer gives agents read-only access to accounts, campaigns, and ads.

Paid search produces a steady flow of questions: which campaigns are burning budget, which keywords are converting, how cost per conversion moved after a bidding change. An AI agent can work through these quickly, but only if it has clean, complete data. The Google Ads API is detailed and powerful, and it is also strict. It has its own query language, retention limits, manager-account rules, and rows that silently disappear when metrics are zero.

Keeping a governed copy of Google Ads data in your warehouse gives agents a stable surface to query. This guide covers how Airbyte builds that copy and what agents can do with it.

Why Google Ads data matters for agentic AI

Google Ads data ties spend to intent. Each search term, keyword, and ad links budget to what a buyer was looking for. For an agent, that makes it useful in several ways:

  • Spend accountability. Cost and conversion data by campaign, ad group, and ad lets an agent locate where money is going and what it returns.
  • Intent signals. Keyword and search-related reports show what people searched for before clicking.
  • Geography and device. Geographic and device segments help explain regional or device-specific changes.
  • Joins with revenue. In a warehouse, Google Ads data can be joined with CRM opportunities or billing data so agents can reason about pipeline and revenue, not just clicks.

What agents can do with Google Ads data

  1. Diagnose cost changes. Compare daily performance by campaign and ad group to find where cost per conversion rose, then list the ads and keywords involved.
  2. Find wasted spend. Surface keywords or ad groups with spend and no conversions over a chosen period.
  3. Compare regions. Use the geographic performance report to break down results by region, metro, or city.
  4. Answer status questions. With agent access on, list accessible accounts, campaigns, ad groups, and ads and check their current status without waiting for a sync.
  5. Run custom analyses. Teams can add custom GAQL queries for the exact fields their reports use, and agents can query those tables like any other.

How Airbyte gets Google Ads data ready

Authentication and access

The connector authenticates with Google OAuth. Since Google sunset developer tokens on September 9, 2026, API access levels are granted to the Google Cloud project that owns your OAuth client. The connector still requires a developer token field, so you can enter an existing token or any non-empty placeholder.

If you access accounts through a Google Ads manager account, enter the manager account's customer ID.

Sync modes

The connector supports Full Refresh Overwrite, Full Refresh Append, Incremental Append, and Incremental Append + Deduped. Incremental streams sync up to the previous day in the account's time zone, because Google Ads filters data by date rather than time.

Streams

The connector syncs Google Ads resources and reports, including:

  • customer, campaign, ad_group, ad_group_ad, labels
  • account_performance_report, ad_performance (ad-level report segmented by date, network, and device)
  • keyword_view, geographic_view_with_metrics, shopping_performance_view
  • click_view, which retrieves one day at a time and only covers the last 90 days

Report streams that include metrics are only available for non-manager accounts.

Custom GAQL queries

You can add custom queries in Google Ads Query Language, each with its own destination table name. If a query selects segments.date, the connector adds it to the WHERE clause and syncs the query one day at a time.

Deletes, conversions, and history

  • Deletions. For ad_group_criterions, ad_listing_group_criterions, and campaign_criterion, incremental syncs read the Change Status resource first and capture deletions. Change Status only covers the most recent three months and returns at most 10,000 records per request. For larger change volumes, Airbyte recommends a full refresh.
  • Conversion window. Conversions can be recorded days after a click. The conversion window setting, 14 days by default, controls how far back the connector re-reads to pick up late conversions.
  • Retention. Report streams and custom queries that use segments.date are limited to Google's 37-month retention window.
  • Zero rows. The API omits rows where all requested metrics are zero, so agents should treat a missing row as zero activity, not missing data.

How agents reach Google Ads data

Google Ads is an agent connector in Airbyte Cloud:

  • Agent access (direct reads). Agents can list accessible customers, accounts, campaigns, ad groups, ad group ads, and labels through the context layer. Reads use the credentials saved in Airbyte, need no sync, and are read-only. They count toward your Google Ads API limits.
  • Search. When a connection syncs Google Ads to an agent-connector destination, the selected streams can be indexed for keyword, semantic, or hybrid search.
  • Warehouse SQL. With BigQuery or Snowflake as the destination, agents can run read-only SQL over the full history, including report streams and custom queries.

Agents connect through the Airbyte MCP at https://mcp.airbyte.com/mcp. Direct reads and searches are free up to 100,000 combined per organization per month. The context layer is available only in Airbyte Cloud.

What this unlocks

  • Paid search analysis from a complete, typed copy instead of live GAQL calls.
  • Late conversions reflected through the conversion window.
  • Custom GAQL tables tailored to your own reporting.
  • Read-only status lookups for accounts, campaigns, and ads.
  • Cross-channel reporting with Facebook, Bing, TikTok, and revenue data.

Getting started

  1. Create a Google Ads source in Airbyte Cloud and sign in with Google.
  2. Enter the manager account ID if needed and add any custom GAQL queries.
  3. Sync to BigQuery or Snowflake with incremental report streams.
  4. Turn on agent access and connect your agent to the Airbyte MCP.

Try Airbyte free or talk to sales about paid media data for agents.

Frequently asked questions

Do I still need a Google Ads developer token?

Google sunset developer tokens on September 9, 2026, and API access levels now belong to the Google Cloud project that owns your OAuth client. The connector still requires the field, so you can enter an existing token or any non-empty placeholder.

Why are some rows missing from Google Ads report streams?

When a query includes metrics, the Google Ads API omits rows where all metrics are zero. Report streams can therefore return fewer rows than dimension-only streams for the same resource.

How much Google Ads history can Airbyte sync?

Report streams and custom queries that use segments.date are limited to Google's 37-month data retention window, and the connector skips anything older. The click_view stream only covers the last 90 days.

Does the Google Ads connector capture deletions?

For ad_group_criterions, ad_listing_group_criterions, and campaign_criterion, incremental syncs use the Change Status resource and capture deletions, but only for the most recent three months.

Can I sync Google Ads data for a manager account?

Yes, enter the manager account's customer ID. Report streams with metrics are only available for non-manager accounts, so select the client accounts you want performance data for.

Integrate with 700+ apps using Airbyte

Move data from 700+ sources into warehouses, lakes, and beyond. Set up pipelines in minutes with pre-built connectors and the Connector Builder.