Loading page…
The pilot accepts both restrictions and preferred wording. Microsoft says it can reduce off-brand copy and editing work, but does not promise that every instruction will be followed.
Listen to this story
Microsoft Advertising’s open pilot gives advertisers a way to set up to 40 messaging guidelines before AI generates campaign copy, including preferred language as well as claims, themes, tone, or concepts to avoid. It is limited to Performance Max and Search campaigns with AI Max enabled, so it is not a platform-wide release. The controls could make brand direction more explicit and reduce manual review, but Microsoft reported no measured gains, and the pilot leaves open how consistently the system follows guidance or resolves conflicts.
Microsoft announced the open pilot on October 28, 2026.
Microsoft’s example guideline for a high-end retailer says not to imply that its products are low cost.
Advertisers enter the guidelines during campaign creation, before the system generates copy.
Advertisers can now give Microsoft’s ad-writing AI a written set of boundaries—not just ask it to produce copy. Microsoft Advertising introduced an open pilot on October 28, 2026, accepting up to 40 messaging guidelines for Performance Max campaigns and Search campaigns with AI Max enabled. The instructions can cover both unwanted messages and preferred language.
The feature, called messaging guidelines, gives advertisers a place to describe how generated ads should sound and which ideas they should avoid. Microsoft recommends using the instructions primarily for broader rules about claims, themes, concepts, tone or language, rather than framing the feature solely as a list of forbidden words.
Its example concerns a high-end retailer that does not want its products presented as inexpensive. The instruction addresses the meaning of the sales pitch, not a particular product description or a finished headline.
Never imply that products are low cost
Microsoft Advertising’s example guideline for a high-end retail brand
The brief can also tell the AI what to favor. Advertisers may supply preferred wording, themes or tone of voice, so the feature is not limited to negative instructions. That gives the same set of guidelines two jobs: keeping certain messages out and encouraging a recognizable way of speaking.
Microsoft shows the guidelines being set during campaign creation. Its stated approach is to establish those boundaries upfront, before the copy is generated. The supported campaign types are specific: Performance Max and Search campaigns that have AI Max enabled. The announcement describes access as an open pilot, not general availability.
Microsoft says the instructions help reduce the risk of generated messaging falling outside a brand’s business requirements. It also says clearer boundaries help maintain brand consistency and reduce time spent reviewing and editing copy. Those are the company’s stated benefits, rather than measured results disclosed with the pilot.
The distinction is important: Microsoft describes the guidelines as a way to steer output, not a guarantee that every ad will comply. PPC Land noted that the announcement provides neither a guideline-violation rate nor an explanation of how conflicting instructions are handled. Advertisers get a more detailed way to state their requirements; how reliably the AI follows them remains unresolved.
Loading discussion...
Join the conversation
Explain whether that disclosure would change your willingness to use them.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.