Ask Advisor finished off the busy work.
Google opened the Ask Advisor beta on May 20. It is a single Gemini-powered agent that crosses Google Ads, Analytics, Merchant Center, and Google Marketing Platform, maintaining context across products 1. Tell it to launch a campaign and it pulls product data from Merchant Center, drafts the creative, sets the bid strategy, and runs a performance report from Analytics afterwards. In one prompt.
The reaction in the industry trade press has been to call this the start of agentic marketing. That framing is correct but slightly behind. Most senior account managers have been doing 60 to 80% of their operational work through GPT, Gemini, or Claude wrappers for the past 18 months. Ask Advisor is not introducing AI to ad-ops; it is putting Google’s version of that workflow inside the platform, with native access to the data the third-party tools cannot see.
The interesting question is what changes for the people who used to do that work by hand.
What Ask Advisor does
Inside Google Ads, Ask Advisor can draft RSA headlines and descriptions, recommend keyword additions, suggest bid strategy changes, build pMax campaigns from a Merchant Center feed, and explain why performance shifted on a given day 2. Inside Analytics, it can run cohort analyses, attribute revenue to channels, and surface anomalies. Inside Merchant Center, it can audit feed quality and flag product-level issues. Inside GMP, it can build audience segments from natural-language descriptions.
The cross-product piece is the part that did not exist before. A reporter at the keynote demonstration asked Ask Advisor why a campaign’s ROAS dropped on Tuesday. The agent pulled the campaign data from Ads, the user behavior from Analytics, the inventory status from Merchant Center, and the audience composition from GMP, then returned a four-sentence answer naming the cause 3. That cross-product diagnosis used to take an analyst 30 minutes.
The agent is in beta for English-language accounts only. It is not auto-applying changes; everything requires confirmation. That will change. The next 12 months will see auto-apply expand into more campaign-management areas, and by late 2027 the default will likely be auto-apply with human override, not the other way around.
What it does not do
Ask Advisor does not invent strategy. It does not know whether your offer is competitive. It does not know whether your landing page matches your ad. It does not know that the ICP you are targeting on LinkedIn is not the ICP that converts on Google. It does not know that the creative you keep running has been fatigued for two weeks.
Those judgments require context that lives outside the ad accounts: in your CRM, in your sales team’s notes, in customer interviews, in the actual product. Ask Advisor will get better at pulling from connected sources, but the strategic decisions that determine whether the account works are not problems Gemini can solve from the data it has.
It also does not catch its own mistakes well. In the firm’s beta testing across three accounts so far, Ask Advisor confidently recommended consolidating ad groups that should have stayed separate, suggested broad match additions that would have absorbed budget into low-intent traffic, and twice mis-read a Merchant Center feed warning as a performance issue. The recommendations are mostly good. The bad ones are presented with the same confidence as the good ones. An operator who accepts without auditing will lose money.
Where the hours go now
The work that Ask Advisor compresses is real. A senior account manager in a traditional agency spends roughly 40 to 60% of their week on tasks the agent can now do faster: pulling reports, drafting copy variants, building audience segments, explaining performance shifts, launching test campaigns. Compressing that to 10 to 20% frees 20 hours per account per month.
Where those hours go is the question that separates the agencies that survive this transition from the ones that do not. The hours have to go to the work the agent cannot do. That work is concentrated in four places.
First, offer design. The strongest single lever in any account is what you are offering, to whom, at what price, with what guarantee. Most accounts the firm audits have not tested an offer variant in six months. Ask Advisor cannot tell you which offer to test next. A senior practitioner who has talked to customers can.
Second, creative concept development. The asset-generation tools (Veo, Imagen, Asset Studio) can produce ten variants of a concept that works. They cannot produce the concept. Concept work is reading buyer language, studying competitors’ tone, watching how the product gets used, and translating that into a hook that lands. That is hours of customer research and creative direction, not prompts.
Third, landing page work. Bid strategy changes produce single-digit performance swings; landing page changes produce two to three times the conversion lift, as the firm laid out in the landing page piece. Ask Advisor will not redesign your form, rewrite your headline, or test a new layout. That is conversion-rate work that requires a human who knows the buyer.
Fourth, attribution and decision-making across paid and organic. Ask Advisor can attribute conversions inside Google’s products. It cannot tell you whether the spend on paid search is cannibalizing organic, whether your branded traffic is being inflated by display retargeting, or whether you should be running paid awareness at all. That is the senior strategic work.
What this means for in-house operators
If you run paid media in-house, the practical shift is to treat Ask Advisor like a senior IC who works for you, not a senior strategist who works above you. Use it to compress the operational tasks. Audit every recommendation. Use the time it saves to do the work that decides whether the account performs.
A useful pattern: open Ask Advisor at the start of your weekly review. Ask it for the three biggest performance shifts of the prior week and the most likely cause of each. Read its answer. Then go look at the actual data. Where it was right, you saved 20 minutes. Where it was wrong, you learned something about the limits of the tool. Over a quarter, that calibrates you to know what to trust and what to question.
The accounts that perform in 2027 will be run by operators who can do the strategy work the agent cannot do. The accounts that underperform will be run by operators who let the agent set the direction.
What this means for agencies
For agencies, this is the most consequential platform change in five years. The traditional agency value proposition, “we manage your ad account so you do not have to,” is in the process of being absorbed by Google. The traditional billing structure, hours times a rate, points the wrong direction: clients will not pay for hours that Ask Advisor compresses to minutes.
The agencies that survive this will look more like the firm’s position on traditional PPC agencies: smaller teams, deeper strategy work, fewer accounts, paid for outcomes rather than for hours. The model that bills $4,000 per month for one analyst to manage four campaigns is over. The model that bills the same for offer design, creative concepting, landing page work, and oversight of an account that runs largely on platform automation is, if anything, more defensible.
The agent is the floor, not the ceiling. The question for everyone working in paid media is what value you add above what Ask Advisor adds for free. If you cannot answer that in one sentence, the next 18 months will be uncomfortable.
The beta starts now
If you have not requested access to the Ask Advisor beta, do that today. The first week you spend with it is the most useful: every prompt you try teaches you what it can do, what it cannot, and what your job is going to look like in 12 months.
The tools change. The leverage does not. The accounts that win are still the accounts with the better offer, the better creative concept, the cleaner feed, the sharper landing page. Ask Advisor is the layer below that work. The work above it is, finally, all that matters.