Go back

Article

Jun 10, 2026

Google's Ask Advisor: What It Does Well, and Whose Side It's On

Ask Advisor is Google's agent operating inside your ad account. Here's what it's good at, where the incentive bends, and the settings to switch off this week

Single thin orange seam between two converging black architectural planes in deep shadow

TL;DR

Ask Advisor, launched at Google Marketing Live 2026 on May 20, spans Google Ads, Analytics, and Merchant Center.
It's strong at diagnostics, draft creatives, and account housekeeping, weaker on bid and budget calls.
The agent is owned by the platform that bills you. That shapes its recommendations.
Keep auto-apply off for budget, bid strategy, and broad-match expansion. Review-first for everything else.
Run a weekly change-log review. Pair it with a second pair of eyes, agent or human.

The short answer

The google ads ask advisor is a useful diagnostic and drafting partner, and a poor unsupervised buyer. It's the agentic layer Google rolled into its ad stack at Google Marketing Live 2026 on May 20, 2026, and it sits inside Google Ads as a product surface alongside Analytics Advisor and Marketing Advisor (Google support documentation). Use it to investigate why a campaign dropped, to draft RSA variants, to catch missing conversion actions. Don't let it apply budget changes, bid-strategy swaps, or broad-match expansions without a human looking at the diff first. The operating procedure is review-before-apply, with specific auto-apply checkboxes disabled. The rest of this piece walks through which ones, and why.

1. From Ads Advisor to Ask Advisor: what shipped and when

Google has been seeding the advisor surface inside Google Ads for over a year. The May 20, 2026 announcement at Google Marketing Live formalised it: a single conversational entry point, Ask Advisor, now spans Google Ads, Analytics, and Merchant Center under one branded layer (Google blog). The earlier per-product advisors, Ads Advisor, Analytics Advisor, Marketing Advisor, remain documented as in-product surfaces (support article).

What changed in practice is the scope of the ask. You can pose a question that reaches across the three properties, "why did conversions drop on Performance Max last week, and is it a feed issue or a bid issue?", and the agent will pull from all three. That's a genuine improvement for gemini in google ads as a diagnostic surface. It does not, by itself, change who the agent works for.

2. Task-by-task: what it genuinely does well

After eight weeks of using the new surface across client accounts, the google ads advisor ai earns its keep in three lanes.

Diagnostics is the strongest. Asking "what changed in the last 7 days on Campaign X" returns a clean attribution of CPC drift, impression-share loss, conversion-rate movement, and feed-side issues for Shopping. It's faster than building the same view manually, and it surfaces the why in plain language. For an operator running 20+ campaigns, that's an honest 30–45 minutes back per diagnostic round in our client work.

3. The incentive problem: an agent owned by the party you pay

Here's the thing every announcement recap skips. Ask Advisor is an agent operating inside the wallet of the customer it advises, owned by the company that bills the wallet. That is not a scandal. It is a structural fact, and it should shape how you wire it in.

Google's revenue moves when your spend moves. The agent's recommendations are, on average, biased toward expansion: broader match types, higher budgets, more campaign types active, more automation surface enabled. Some of those recommendations are good for you. Some are good for Google. Most are both, partially. None of them are neutral. The agent is doing what it was built to do.

4. Auto-apply recommendations: which to switch off and why

Google Ads ships with an auto-apply recommendations panel that lets the platform implement suggestions without human review. The default settings are not the safe settings.

Disable these immediately:

Raise budgets for limited campaigns. This is the single most expensive auto-apply. Budget limits exist for a reason, usually a margin reason the agent doesn't know.
Add broad-match keywords. Broad match auto-expansion is how accounts quietly bleed into irrelevant traffic. Review-first only.
Switch bid strategies (to Maximise Conversions, Maximise Conversion Value, or tCPA changes). A bid-strategy switch can move spend pattern 20–40% within a week in our client work. Never automatic.
Upgrade to Performance Max from Standard Shopping or Search. A genuine campaign-type migration deserves a meeting, not a checkbox.

If you want a deeper take on the guardrail layer, we wrote one here: Google Ads AI guardrails.

5. A safe operating procedure: review-before-apply, change logs, weekly checks

An operator-grade procedure for ads advisor recommendations safe usage looks like this:

Daily (5 minutes). Open Ask Advisor with one diagnostic question for the campaigns that matter most. Read the answer. Apply nothing. Note anything that looks like a real signal in a shared doc.

The non-negotiable: there must be a change log, and someone must read it. An agent without a change log is an agent without accountability, and the Sinch survey of 2,500+ enterprises found that 75% of companies rolled back customer-facing AI agents, with no diagnostics cited by 16% as a top cause. That's the failure mode. We wrote about the broader pattern here: why companies are rolling back AI agents.

6. When an independent second agent (or human) is worth it

A second supervisory layer earns its cost when monthly ad spend crosses roughly $50K, or when the account has multiple campaign types running, or when the business model has narrow margins where a 15% spend-pattern shift hurts.

That second layer can be a human media buyer, or it can be an independent agent, one whose objective function is your margin, not the platform's revenue. The point isn't to distrust gemini in google ads. The point is that no agent should be its own auditor, regardless of who built it. If you'd like a read on whether your account is at the size where independent supervision pays for itself, that's the conversation we have on the paid ads service page.

At small spend levels, Ask Advisor plus a disciplined weekly review is enough. At scale, it isn't.

7. FAQ

What data does Ask Advisor have access to inside my account?

Ask Advisor reads your Google Ads campaign data, and, when linked, your Google Analytics 4 property and Merchant Center feeds, per the Google Marketing Live 2026 announcement. It does not read your CRM, your billing system, or any data source you haven't explicitly connected. Treat connected properties as the agent's full visibility.

What account permissions does it need, and can I limit them?

Ask Advisor inherits the permissions of the Google account using it. A user with admin access enables broader auto-apply; a user with standard or read-only access cannot. The practical guardrail is to keep auto-apply settings configured at the manager-account level, and restrict admin access to one or two people who own the change-log review.

Can I roll back changes Ask Advisor applies automatically?

Yes. Google Ads keeps a change history for the last 90 days, and applied recommendations appear there with a clear source label. You can revert individual changes from the change-history view. The harder problem is noticing the change happened, which is why the weekly review exists, not because rollback is technically difficult.

Is it safe to let Ask Advisor manage bid strategies on smaller campaigns?

In our client work, no, not without review. Bid-strategy changes look small in the UI and move spend patterns 20–40% within a week. The risk is the same on a $2K campaign and a $200K campaign; only the absolute dollars differ. Keep bid-strategy auto-apply off across all campaigns.

How is Ask Advisor different from the older Ads Advisor surface?

The older Ads Advisor lived inside Google Ads only, documented in the Google support article. The May 20, 2026 launch unified it with Analytics Advisor and Marketing Advisor under one conversational entry point that reaches across all three products. The underlying recommendations are similar; the cross-product visibility is the real change.

What to do this week

Open your auto-apply recommendations panel today. Switch off budget raises, broad-match expansion, bid-strategy changes, and Performance Max upgrades. Put a 30-minute weekly slot on the calendar to walk the change log. That's the whole procedure, and it costs you nothing.

If you'd like a second pair of eyes on the account before you trust the agent any further, we're here.

FAQ

How should a small team prioritize google ads?

Start with the workflow that already has a baseline: hours, leads, errors, or budget waste.

What should be measured before investing in google ads?

Measure cycle time, volume, handoffs, error rate, and the current owner.

When should google ads ask advisor stay manual instead of automated?

Keep it manual when judgment, approval, brand nuance, or customer trust is on the line.

How does google ad change the budget for google ads?

google ad usually adds integration, QA, and monitoring work.

What is the first project to launch from this google ads ask advisor playbook?

Launch the narrowest workflow with a visible result.

How often should google ads be reviewed?

Review it weekly for the first month, then monthly once stable.

What mistakes make google ads ask advisor fail?

The common mistakes are starting too broad, skipping measurement, granting write access too early, and not naming an owner.

Can google ads work without a full rebuild?

Yes.

What data does google ads ask advisor need before it can improve results?

It needs clean inputs, timestamps, outcome labels, and a correction loop.

Who should own google ads after launch?

Give ownership to the person closest to the outcome.