Microsoft Advertising rolled its own AI Max tool to general availability in September 2026, enabling it by default for all new Search campaigns. Alongside the rollout, Microsoft reported a conversion-lift figure with more specificity than most platform announcements bother with: a lower confidence bound of 8.2% conversion lift, with a spend-weighted lift around 13.6%, across 44 advertiser A/B tests. If you're running Microsoft Ads and haven't checked whether AI Max is now on by default for your new campaigns, that's worth confirming this week. "On by default" means it applies whether or not anyone on your team opted in.
Here's the detail worth noting before treating that lift number as your own expected outcome. A spend-weighted average across 44 tests tells you what happened in aggregate, weighted toward the accounts that spent the most in the sample. It doesn't tell you where your account, at your spend level and in your vertical, would land inside that range. Microsoft naming a confidence bound at all is more rigorous than a bare average would be, but it's still one platform's own reported test set, not an independent benchmark.
- Microsoft's AI Max reached general availability in September 2026 and is enabled by default for new Search campaigns.
- Reported lift: a lower confidence bound of 8.2%, with spend-weighted lift around 13.6%, across 44 advertiser A/B tests, Microsoft's own reported figure.
- This lands the same month Google's own AI Max is force-migrating a different set of Search campaigns, which is a naming coincidence worth clarifying, not a shared product.
- "On by default" for new campaigns means the setting applies without an opt-in step, which is the detail most likely to be missed by a team not actively watching Microsoft Advertising's changelog.
- Microsoft Ads remains the platform with the thinnest content and community coverage of the six major ad platforms, which is exactly why concrete numbers like this are worth a dedicated look rather than a footnote.
Is Microsoft's AI Max the same product as Google's AI Max?
No. Same name, different platforms, different products, and the naming collision is going to cause real confusion this quarter given both launched AI Max-branded features in the same general window. Microsoft's AI Max is Microsoft Advertising's own tool for expanding Search campaign reach and improving audience targeting through AI-generated personalization and expanded query matching. Google's AI Max, covered separately, is Google Ads' equivalent concept, also focused on Search, also about expanded matching and AI-assembled assets, but it is a distinct product on a distinct platform, with its own settings, rollout mechanics, and now its own forced-migration timeline.
If your team manages both Google Ads and Microsoft Ads, the practical risk is talking about "AI Max" in a planning meeting and having two people mean two different platform features. Name the platform every time this comes up internally until the naming settles, because it won't settle on its own.
What does Microsoft's AI Max actually do?
Two components, similarly structured to what Google calls AI Max even though the products are separate: expanded query matching that finds more of the relevant query space around your existing keywords, and an audience-generation capability that lets advertisers describe an ideal customer in plain language, which the system translates into a targetable audience segment. Together, the pitch is broader reach with less manual keyword and audience-list maintenance.
The audience-generation piece is the newer and less-tested half of the two. Expanded query matching is a variation on something search platforms have done for years; describing an audience in natural language and having the system build the segment is a genuinely different workflow, and it's worth treating the segments it generates as a hypothesis to validate against your own conversion data in the first few weeks, not as a ready-made replacement for a manually built audience list you already trust.
How should you read the 8.2%-13.6% lift figure?
Read the two numbers as bounding a range, not as one figure. The 8.2% is described as a lower confidence bound, a statistically conservative floor, the number Microsoft is confident the true effect is at least as large as, across the tested population. The 13.6% is a spend-weighted average: the mean lift, but weighted so that higher-spending advertisers in the sample count for more of the average than lower-spending ones.
That distinction matters for a specific reason: a spend-weighted average can be pulled upward by a small number of large accounts that saw an outsized effect, while the lower confidence bound is a more conservative, broadly-applicable statement. If your account looks more like the median advertiser in the test set than the highest-spending ones, the 8.2% floor is the more honest number to anchor your own expectations on, not the 13.6% headline figure that's more likely to lead a press summary.
What neither number tells you: how the 44 tested advertisers were selected, across what verticals, or over what time window, details that would meaningfully change how confidently you'd extrapolate to your own account. Treat this as a real, above-average reported result from the platform running the test, not as an independently audited benchmark.
Why does a Microsoft Ads-specific update deserve its own post?
Because Microsoft Advertising is genuinely underserved in the current wave of platform-AI coverage, and that gap has a real cost for teams running it. Google Ads and Meta dominate the AI-in-advertising conversation by volume of content and community discussion, which means Microsoft-specific product changes, including ones with real, named performance data attached, like this one, get less scrutiny and slower community validation than equivalent Google or Meta announcements.
For a team running Microsoft Ads as one of several platforms, that's a reason to treat Microsoft's own release notes as a primary source you check directly, rather than waiting for the update to surface through the same channels that cover Google and Meta more thoroughly. It's also a reason a change like "on by default for new campaigns" is more likely to be missed here than on a more heavily-watched platform.
What should a team running Microsoft Ads check this month?
Confirm AI Max's default status on your new campaigns. General availability with a default-on setting means new campaigns get it whether or not anyone requested it, worth a quick account check rather than an assumption either way.
Validate audience-generation output against known-good segments. Before replacing a manually-built, historically strong audience list with an AI-generated one, run them in parallel long enough to see whether the generated segment's conversion behavior actually matches or beats the one you trust.
Don't import Google's AI Max playbook wholesale. The products share a name and a general concept, broader Search matching, less manual maintenance, but the underlying mechanics, rollout timeline, and reported results are specific to each platform. What you learn auditing one doesn't transfer automatically to the other.
Frequently asked questions
Related reading
- Microsoft Ads + AI Agents: The Complete Guide: how agent-run execution fits alongside Microsoft's own native AI features
- Microsoft Ads is now live on Adlyse: Adlyse's own Microsoft Ads coverage
