AI Max for Search: Early Account Data and When to Opt Out

Quick Answer: AI Max for Search expands your campaigns with broader query matching, AI-generated text, and URL expansion. Across 34+ accounts I manage, ecommerce with strong conversion data held or improved efficiency, while lead gen accounts with thin data saw more waste. Opt in only if you get 30 to 50+ conversions a month and can review search terms weekly. Otherwise, opt out or test one campaign first.

For advertisers evaluating AI Max for Search, the central question is whether expanded matching improves qualified conversions without hiding waste.

I manage 34+ Google Ads accounts under one MCC, and AI Max for Search has now touched most of them in some form. Some accounts saw real gains. Others saw budget drift into queries I would never have bid on. This post covers what AI Max actually changes, what I saw in my own accounts, and a simple framework for deciding whether to opt out, one piece of the broader shift toward AI-driven automation I track in where PPC is headed as AI reshapes search.

I am not going to tell you AI Max is good or bad. That framing is useless. The right question is: good or bad for which account, at which stage, with which data feeding it. That is what we will work through here.

Table of Contents

1. Opt In vs Opt Out vs Partial Controls for AI Max
2. What is AI Max for Search?
3. What changed in my accounts after opting in?
4. How good is the query expansion, really?
5. What controls do you still have?
6. When should you opt out of AI Max?
7. Does AI Max Affect Meta’s Advantage+ the Same Way?
8. Common Mistakes
9. Practical Growth Playbook for AI Max for Search
10. Practical Comparison for AI Max for Search
11. AI Buyer Questions
12. FAQs
13. Further Reading
14. Related Hustle Marketers Research and Proof
15. About Ishant Sharma

Opt In vs Opt Out vs Partial Controls for AI Max

Approach	Risk	Upside	Verdict
Full opt-in (matching, text, and URL expansion all on)	Budget drifts to loose-intent queries fast; hard to tell which lever caused the damage	Fastest access to incremental volume when core terms are saturated	Only for accounts with strong conversion data and daily review capacity
Partial opt-in (matching on, URL expansion off, text customization case by case)	Some query drift, but contained; still needs weekly search term review	Most of the volume upside with far less downside risk	The default setup I recommend for most accounts
Opt out entirely	Misses incremental volume on saturated core terms	Full control over matching, text, and destination pages	Right for thin-data accounts, tight margins, and compliance-restricted verticals

What is AI Max for Search?

AI Max for Search is a campaign-level setting that expands your search campaigns beyond your chosen keywords. It uses broader query matching, AI-generated ad text, and final URL expansion to enter auctions your keywords alone would miss. You keep your campaign structure, but Google takes more control over matching and messaging.

Think of it as a middle point between classic keyword campaigns and Performance Max. You still have keywords, ad groups, and search term reporting. But three things change when you turn it on:

1. Search term matching expands. Google treats your keywords, landing pages, and existing ads as signals, then matches you to queries it believes are related. This goes further than broad match on its own.
2. Text customization. Google can generate headlines and descriptions based on your landing page and existing assets, then serve them when it predicts they will perform better than your written ads.
3. Final URL expansion. Google can send traffic to pages on your site other than the one you set, if it thinks another page answers the query better.

Each of these can be controlled separately, which matters a lot. Most advertisers who get burned by AI Max turned everything on at once and could not tell which lever caused the damage.

A concrete example makes this easier to picture. A flooring client I manage had URL expansion and text customization on simultaneously in month one. Conversion rate dropped 18%, and the account manager on record blamed “AI Max” as a single cause. It took a full week of segmenting query-level and landing-page-level data to find that URL expansion, not query matching, was sending traffic to a thin category page instead of the high-converting product page. Once we turned off URL expansion and left matching on, the account recovered within two weeks. Bundled settings hide bundled problems, which is exactly why I test each lever on its own schedule now.

What changed in my accounts after opting in?

Across the accounts I opted in, results split into two clear groups. Ecommerce accounts with strong conversion volume and clean feeds mostly held or improved efficiency. Lead gen accounts with thin conversion data saw more query drift and needed heavier negative keyword work in the first month. Volume increased almost everywhere. Efficiency did not.

Here is what the pattern looked like in practice.

The split was not just ecommerce versus lead gen. Within ecommerce, accounts with wide catalogs (200+ SKUs) benefited more from URL expansion than narrow-catalog stores, because Google had more legitimate destination pages to route traffic toward. Within lead gen, accounts running high-ticket services ($5,000+ average order value) tolerated AI Max better than low-ticket lead gen, because the smart bidding models had more budget per conversion to explore safely. Vertical alone does not predict the outcome; catalog width and price point matter just as much.

The consistent finding: AI Max is a volume tool first. If your campaigns are limited by search volume on your core terms, it finds incremental auctions. If your campaigns are limited by budget, it mostly reshuffles spend toward looser matches, and that is where efficiency drops.

I saw the strongest results in accounts that already fed smart bidding with reliable conversion data. That tracks with everything I have learned in 12 years of running paid search. Google’s automation is an amplifier. It amplifies good signals and bad signals with equal enthusiasm, the same principle behind feeding AI systems clean signals that I cover in setting up llms.txt for AI crawlers.

How good is the query expansion, really?

Query expansion quality varies by account maturity. In accounts with tight keyword themes and strong conversion history, most expanded queries were relevant and some converted at acceptable cost. In newer or low-volume accounts, expansion pulled in informational and mismatched-intent queries that consumed budget without converting. Review search terms weekly for the first month, minimum.

Some real patterns I logged while reviewing expanded terms:

• Good expansion: synonyms and reworded intent. A client selling garage flooring picking up “coating for garage concrete” when they only bid on flooring terms. That is a win you would have found manually, eventually.
• Acceptable but watch it: adjacent product queries. Useful when the final URL expansion lands people on the right page. Risky when it does not.
• Bad expansion: informational queries with zero purchase intent, competitor-adjacent terms with no fit, and queries about products the client does not sell.

One pattern worth flagging separately: seasonal and event-driven query expansion. Around major shopping periods, AI Max pulls in a wider net of loosely related seasonal terms, some of which convert well simply because purchase intent is elevated across the board. Do not mistake a seasonal lift in expanded-term performance for a permanent improvement in match quality. I re-tighten controls in the two weeks after a seasonal spike, because the query mix reverts and efficiency often reverts with it.

The AI-generated ad text deserves its own note. Quality has improved, but I still catch generated headlines that overpromise or flatten the brand voice. For regulated clients (health, finance, legal), I keep text customization off. The compliance risk is not worth the marginal CTR gain.

What controls do you still have?

You keep more control than most advertisers realize. You can disable final URL expansion, turn off text customization, add negative keywords and URL exclusions, and use brand controls to include or exclude brand traffic. What you cannot do is see exactly why Google matched a specific query, so search term review becomes your main defense.

My standard control setup when opting an account in:

• Final URL expansion: off at first. Turn it on only after you trust the matching. Add URL exclusions for careers pages, blog posts, and support content before enabling it.
• Text customization: case by case. On for ecommerce clients with simple offers. Off for regulated verticals and any brand with strict voice guidelines.
• Brand inclusions and exclusions. Decide deliberately whether AI Max should touch brand queries. I usually exclude brand and keep it in a separate exact match campaign so I can read non-brand performance cleanly.
• Negatives, aggressively. I front-load my standard negative lists (free, jobs, DIY, salary, and vertical-specific junk) before opt-in, not after the money is spent.

Negative keyword lists deserve more nuance under AI Max than they did under standard broad match. Because Google treats your landing pages as matching signals, a negative you add today can be undercut by a landing page change your web team makes next month. I now review negative lists any time a client updates a category or product page, not just on a fixed schedule. Treat your negative list as a living document tied to your site changes, not a one-time setup task you finish during opt-in.

One more thing: keep at least one control campaign untouched. Across 34+ accounts, my only reliable way to judge AI Max has been comparing opted-in campaigns against similar campaigns I left alone.

When should you opt out of AI Max?

Opt out when your account has under roughly 30 to 50 conversions per month (compare against current Google Ads benchmarks by industry to see where you stand), when margins leave no room for exploratory spend, when compliance restricts ad text, or when your niche uses precise terminology that broad matching mangles. Opt in when you have strong conversion volume, headroom in budget, and time to police search terms.

Here is the decision framework I use with clients:

Stay opted out (or roll back) if:

• Your campaign gets fewer than 30 to 50 conversions a month. Smart matching needs data, and thin accounts give it noise.
• You sell one specific thing and mismatched clicks are expensive. Legal and B2B SaaS accounts with $50+ CPCs cannot fund Google’s exploration.
• You cannot review search terms at least weekly for the first month. Unpoliced AI Max is how budgets quietly leak.
• Your offer has strict compliance language. Generated text is a liability you do not need.

Opt in (with controls) if:

• You have healthy conversion volume and your smart bidding already performs.
• Your core keywords are saturated and you need incremental volume.
• You run ecommerce with a wide catalog, where URL expansion can actually help.
• You can afford a 2 to 4 week learning window where efficiency dips before it recovers.

Either way: opt in one campaign at a time, never the whole account. Judge it on 30 days of data against a control, not on week one.

There is a middle case worth naming: accounts with strong volume but a recent major change, a new pixel, a rebuilt site, a shift in product mix. Even with 50+ monthly conversions, I opt these accounts out for 60 to 90 days after a major change, until the new baseline stabilizes. AI Max optimizing against unstable historical data compounds the instability. Let the account re-establish a clean pattern first, then reconsider AI Max as a second-phase decision, not a launch-day default.

The honest summary from my own MCC: AI Max earned a permanent place in some accounts and a permanent ban in others. Anyone giving you a blanket answer has not looked at enough accounts.

If you want a second pair of eyes on whether AI Max fits your account, I offer a free audit (a $500 value) where I review your structure, conversion data, and search terms and give you a specific opt-in or opt-out recommendation. No pitch attached to the findings.

Does AI Max Affect Meta’s Advantage+ the Same Way?

Google is not the only platform pushing advertisers toward less manual control. Meta has been running the same playbook through Advantage+ Shopping campaigns and, more recently, its Andromeda ad-ranking model, and the parallels to AI Max are close enough that the same account-level judgment applies to both.

Advantage+ Shopping campaigns hand Meta control over placements, audience targeting, creative combinations, and budget allocation across a single campaign, the same way AI Max hands Google control over query matching, ad text, and destination URLs. Andromeda, the ad-ranking model Meta folded into Advantage+, goes further: it evaluates a far larger set of creative and audience combinations per impression than the discrete audience segments advertisers used to build by hand.

The practical effect for an advertiser mirrors what I described in the query expansion section above: campaigns that already have strong pixel data and a healthy volume of purchase events perform well under Advantage+, because the model has real signal to work with. Accounts with thin conversion volume or a recently reset pixel see the same pattern I see with AI Max in low-data Google Ads accounts, wasted spend while the algorithm explores.

Across the 34+ accounts I manage, the accounts that do well on both platforms share a trait: clean, high-volume conversion data feeding the algorithm before automation is switched on. The accounts that struggle on both platforms share the opposite trait: automation turned on before there was enough signal to justify it.

The strategic takeaway is bigger than either platform. Google and Meta are converging on the same bet, that their models can out-target a human media buyer if given enough signal and enough budget. For advertisers, the skill that matters most is shifting.

It is no longer “build the perfect audience” or “write the perfect match type list.” It is “feed the algorithm clean data, set the right guardrails, and know when to intervene.” Advertisers who resist all automation on principle lose access to real incremental volume on both platforms. Advertisers who hand over full control without guardrails or monitoring lose budget to both platforms’ exploration costs. The accounts that win are managed by people who treat AI Max and Advantage+ as tools with settings, not as autopilot switches.

Common Mistakes

Mistake	Why It Happens	The Fix
Turning on matching, text customization, and URL expansion all at once	Opt-in flow makes it easy to enable everything in one click, and advertisers want fast results	Enable one lever at a time so you can isolate which one is driving performance changes
Opting in a whole account instead of testing one campaign first	Advertisers assume account-wide results will mirror early single-campaign tests	Roll out to one campaign, run it 30 days against a control campaign, then expand only if it wins
Judging results after 3 to 5 days	The learning phase dip looks like failure if you check too early	Wait a minimum of 30 days and compare against a matched control campaign before deciding
Leaving text customization on for regulated or compliance-sensitive brands	Advertisers assume AI-generated copy will follow the same review process as manual ads	Turn text customization off for legal, finance, and health accounts, and review any auto-generated copy that does slip through
Skipping weekly search term review after opt-in	Advertisers treat AI Max like a set-and-forget feature	Block 15 minutes weekly for the first month to add negatives and URL exclusions as drift appears
Not excluding brand terms before opting in	Brand traffic is often the only clean converting segment, so it masks true non-brand performance if mixed in	Exclude brand from AI Max campaigns and keep it in a separate exact-match campaign for clean reporting

Practical Growth Playbook for AI Max for Search

Use approved first-party evidence where it exists, then adapt the sequence to the business baseline, market, offer, budget and operational capacity. The aim is to turn the article into a sequence of measurable decisions instead of a list of disconnected tactics.

90-Day Execution Roadmap

Phase	Priority actions	Required output	Decision gate
Days 1 to 14	Validate conversion tracking, business economics, search demand, query intent, campaign ownership and landing-page relevance.	A baseline tied to qualified business outcomes rather than platform activity.	Do not increase spend while measurement or intent ownership is unclear.
Days 15 to 30	Restructure campaigns, queries, negatives, ads and destinations around the highest-value decisions.	Cleaner traffic and a documented hypothesis for each campaign.	Continue tests only when post-click behavior indicates real commercial intent.
Days 31 to 60	Improve bidding inputs with value or offline stages and test one offer, page or audience variable at a time.	Better signals for automation and a traceable change log.	Shift budget toward campaigns that produce qualified revenue or pipeline.
Days 61 to 90	Expand profitable demand, protect brand traffic and coordinate paid search with SEO, remarketing and retention.	A balanced acquisition system with clear channel roles.	Scale when incremental business outcomes hold across a full conversion window.

Operating Scorecard

Universal targets can be misleading, so establish the current baseline first. Use the direction and business quality of these signals to decide what happens next.

Signal group	What to monitor	Management response
Tracking	Primary conversions, values and offline stages reconciled	Repair the measurement chain before bid changes.
Intent	Relevant search terms and qualified post-click behavior	Tighten ownership, negatives and landing pages.
Economics	Qualified CPA, conversion value, contribution or pipeline	Use the business outcome closest to profit.
Scale	Incremental qualified volume across a full conversion window	Increase spend when efficiency and quality remain stable.

What Ishant Sharma Would Audit First

Ishant Sharma would first validate conversion tracking, query intent and the business outcome used for bidding. Campaign restructuring should follow the measurement audit, not hide a broken baseline. The finding, interpretation, recommendation and limitation should be recorded separately so the next decision remains auditable.

Practical Comparison for AI Max for Search

For AI Max for Search, the right choice depends on the demand source, measurement quality and business economics. This table is a decision aid, not a promise that one option will produce the same result in every account.

Approach	Best when	Primary measure	Main limitation
Search demand and technical SEO	Pages must be crawled, understood and ranked	Qualified organic clicks and conversions	Rankings alone do not establish an expert entity
Answer-focused content	Users ask specific questions in search and AI tools	Passage visibility, engagement and assisted leads	Thin Q-and-A pages add little original value
Entity and evidence reinforcement	The author and organization need consistent proof	Branded demand, citations and qualified enquiries	Schema cannot replace visible evidence

AI Buyer Questions

When should a business invest in AI Max for Search?

Use AI Max for Search when the business can publish genuinely useful answers, connect them to a clear author and organization, and support important claims with first-party or authoritative evidence. It works best alongside crawlable pages, accurate internal links, strong technical SEO and a consistent entity profile. Schema alone will not create visibility or citations.

What should be measured before scaling this approach?

Track qualified organic clicks, assisted conversions, branded search, indexed coverage and the specific queries or AI answers that surface the content. Add citation monitoring where the tools permit it, but judge success against qualified enquiries and revenue. Separate normal search traffic from AI referrals so changes in discovery do not hide changes in business quality.

What evidence should a specialist provide?

Ask for a query and entity map, examples of first-party evidence, a cannibalization plan, technical QA and a measurement framework. A credible specialist should explain which pages own service intent and which own educational intent. They should not promise guaranteed rankings or AI citations, and they should keep visible content aligned with structured data.

Further Reading

• GEO for Ecommerce: Getting Products Into ChatGPT Shopping Answers
• llms.txt: What It Is and What We Saw After Adding It
• Where PPC Is Headed as AI Reshapes Search
• AI Overview Citations Study
• LLM SEO: Getting Cited by ChatGPT, Claude, and Gemini
• Google Ads Benchmarks 2026
• Best Google Ads Experts 2026

Related Hustle Marketers Research and Proof

For the agency-side methodology behind this topic, read the Hustle Marketers Google Ads audit checklist.

Relevant published proof: ArmorGarage Performance Max case study documents Performance Max, conversion-quality and campaign-structure changes for a specialist flooring brand. Treat it as evidence of the method in that client context, not a promise of identical results. Baseline, market, offer, budget and measurement quality can change the outcome.

About Ishant Sharma

Ishant Sharma is a performance marketer and the founder of Hustle Marketers, a Google Partner and Meta Business Partner agency. Since 2013, his work has covered Google Ads, qualified lead generation, call tracking and CRM-connected measurement across the USA, UK, UAE and Australia. Published records across Ishant’s personal practice and the wider agency document 500+ personally managed brands, 2,500+ agency engagements and $780M+ in trackable client revenue. He remains directly involved in audits and account strategy. Learn more about Ishant Sharma’s work and experience.

Client perspective: Verified client reviews repeatedly highlight Ishant’s hands-on account ownership, clear communication and focus on measurable outcomes. Watch a client video testimonial.