Microsoft Advertising opened AI Max to every account: what changes for beauty brands?

AI Max sits on top of existing Search campaigns as a single master toggle with three checkboxes beneath it. Each governs a different stage of the auction (Search Engine Land, 2026; Search Engine Journal, 2026).

• Search term matching. Widens the pool of queries a campaign can enter beyond your keyword list, drawing on keywords, ads, landing pages, user intent and contextual signals. Microsoft positions this specifically for long, conversational queries inside AI experiences on Bing and Copilot.
• Text customization. Generates additional messaging variations from your existing assets and website content, then selects combinations at auction time.
• Final URL expansion. Rather than sending every click to a fixed page, it routes visitors to the page the system judges closest to their intent — conditionally: the interface states the redirect happens when it is "likely to improve ad performance" (PPC Land, 2026).

Controls arrive with the product, not after it

The most notable aspect of this release is that the control layer ships alongside the feature. Brand inclusion and exclusion lists, term exclusions for text asset generation, URL rules, ad group-level settings and optimization experiments capable of testing each feature individually are all available at launch (Search Engine Journal, 2026).

We heard you that brand controls and reporting are critical to trusting AI as a partner in campaign management, which is why brand controls (including term exclusions for text asset generation), reporting, and choice are present from day one.

— Navah Hopkins, Microsoft Ads Liaison, 19 August 2026

There is a numeric cap: up to 20 brand lists can be attached to a single campaign. Worth reading precisely — that is a limit on lists, not on brands, and the interface does not disclose how many entries a list can hold (PPC Land, 2026).

Which campaigns will find the setting already on?

Microsoft is folding two legacy features under the AI Max umbrella: Predictive matching and autogenerated text assets. Campaigns using either will have the corresponding AI Max setting enabled automatically. The rest of the suite does not activate unless the advertiser opts in (Search Engine Land, 2026).

If you import campaigns from Google Ads, AI Max settings carry across. One exception: AI Max campaigns that originated as upgraded Dynamic Search Ads (DSA) are converted back into DSA campaigns on Microsoft while the company builds additional functionality (Search Engine Journal, 2026).

The two products share a name and the same three mechanism names. The divergence is not in the mechanism but in compulsion and priority.

1. No forced migration

Google auto-upgrades campaigns using automatically created assets (ACA) and the campaign-level broad match setting to AI Max on 1 September 2026 — we covered that in detail in our piece on Google Ads' 1 September AI Max migration. Microsoft has announced no conversion date at all; AI Max remains opt-in for new and existing campaigns (Search Engine Land, 2026).

That should not be assumed permanent. Predictive matching ceasing to exist as a standalone targeting option and surviving only as a component of AI Max is a signal about the direction of travel (PPC Land, 2026).

2. Exact match keywords hold absolute priority

On 18 December 2025, Microsoft confirmed that exact match keywords hold absolute auction priority over all other campaign types competing for the same placement. An AI Max expansion inside a Microsoft account therefore sits behind any qualifying exact match keyword. No equivalent guarantee exists on Google's side (PPC Land, 2026).

In cosmetics the practical translation is this: if you hold shade names, product names and brand terms as exact match because you do not want to cede control of them, that protection is structurally stronger on Microsoft.

3. Control limits are not equivalent

Terminology parity conceals mechanical difference. Google's text guidelines cap term exclusions at 25 per campaign; Microsoft's cap operates on 20 brand lists. Cross-platform reporting built on the assumption of equivalence will produce comparisons that are not like for like (PPC Land, 2026).

Of the three, the one that matters most for regulation is text customization, because its output is itself a claim.

Türkiye's medicines and medical devices agency (TİTCK) defines a cosmetic claim as the information, statements or visuals used on the product's "label, packaging and/or promotional materials" (TİTCK, 2024). A search ad headline falls inside that definition. The guideline sets out six common criteria: legal compliance, truthfulness, evidential support, honesty, fairness, and informed decision-making.

The criteria generative copy pushes against by design

• Honesty. The guideline rejects superlatives such as "unique, unmatched, best, most effective" for cosmetic products sharing the same properties. A model writing headlines for click performance gravitates toward exactly those patterns.
• Truthfulness. The guideline's own example is explicit: a product with only short-term moisturising proven cannot claim "provides 48-hour moisturisation." If a duration or percentage on your landing page migrates into a headline, the evidential burden is yours.
• Fairness. The guideline disallows "paraben-free" on the grounds that it disparages the whole paraben group, which is safe when used in line with the regulation. If that phrasing sits on your site, it can surface in generated copy.

The guideline is explicit about who carries the burden: the responsible person must ensure the claim is consistent with the documentation they hold. Even when a model writes the line, the claim is the brand's and so is the proof.

Final URL expansion carries a measurement risk rather than a regulatory one. Destination selection moves from the campaign builder to the platform, which reshapes landing page analytics and conversion attribution. On the Google side this setting has been documented producing 404s and misrouted traffic where tracking templates were non-standard; Microsoft has published no equivalent caveat for its own implementation (PPC Land, 2026).

Search term matching, meanwhile, affects the budget more than compliance: it widens the eligible query pool, not the budget funding it.

First, the prerequisite: Microsoft Advertising is available in Türkiye. The platform launched there in February 2022 as part of a 29-country expansion, which also added Turkish to its supported languages (Search Engine Land, 2022).

Second, Microsoft's actual argument is not Bing volume but conversational queries inside Copilot. Keyword lists were built for a lexical matching era and do not map cleanly onto sentence-length prompts. Search term matching is the bridge Microsoft proposes for that gap (PPC Land, 2026).

In beauty, most pre-purchase questions are already phrased as sentences — "which foundation suits combination skin", "will this lipstick crack on dry lips". Those belong to AI visibility (GEO) rather than to keyword tables; AI Max claims to bring them into paid search.

Third, and most importantly: there is no independent performance measurement at all. Microsoft has not published a comparative performance figure for its own AI Max, and because the product was confined to a pilot population until now, no outside measurement exists either (PPC Land, 2026).

The closest available data point concerns Google's implementation and is not encouraging: a Lunio study published on 12 August 2026, covering more than 414 million clicks recorded between October 2025 and June 2026, reported that Google retail search campaigns running AI Max were exposed to 72% more invalid traffic than those without it. It is worth stressing that this does not measure Microsoft's implementation — the two share a name and a feature taxonomy, not an auction or an inventory pool (PPC Land, 2026).

Every step below maps to a control Microsoft shipped at launch. None of it is guesswork about your account.

• Check what is already on. Campaigns using Predictive matching or autogenerated text assets may already have the corresponding AI Max setting enabled.
• Build your brand lists. A campaign with no list attached inherits no exclusions; the interface flags this with a separate advisory panel. The cap is 20 lists per campaign.
• Write term exclusions in Turkish. Phrases to keep out of generated text assets — "paraben-free", "treats dark spots", "most effective" — have to be listed in the language you advertise in.
• Clean up landing page copy. That is the source material. A risky claim on the site is a claim that can migrate into an ad headline.
• Treat final URL expansion separately. If you enable it, validate tracking templates first and add any page that must never receive paid traffic to the URL exclusion list.
• Do not switch all three on at once. Optimization experiments allow testing features individually. Enable all three together and a single performance movement cannot be attributed to a specific mechanism afterwards.
• Put asset reporting on a schedule. Reading generated headlines regularly is not a set-and-forget task; compliance review only happens by eye.

One date to note: Microsoft Advertising retires its legacy SOAP API on 31 January 2027, with new features exclusive to the REST API from 1 October 2026. Tools managing AI Max controls programmatically through SOAP will lose access to the new settings after that (PPC Land, 2026).