This month’s Ask PPC question about AI ad placements is close to my heart. We saw ads start appearing on AI surfaces since 2024and yet, they still have an air of mystery:
“Ads are starting to appear in AI chat experiences. How should advertisers think about these new placements – and are they worth their budget?”
As I work for Microsoft, I cannot comment on the value for money of competing brands. What I can do is talk about AI ads in general terms:
- How to access AI advertising inventory.
- How to think about metrics for AI investments.
- Build a budget (time and money) for AI internships.
How to access AI advertising inventory
There are actually two ways to purchase AI ad inventory: directly through an AI-driven platform, or as part of your broader paid media buys across major ad networks. Neither is inherently better or worse than the other, but they require different strategies.
If you’re buying direct, you know the media buy is 100% allocated to that AI surface area, making it easier to design creative and measure with that specific experience in mind. These purchases are often available on a cost per thousand (CPM) or cost per click (CPC) basis, depending on the platform and market.
Conversely, when accessing AI surfaces through existing campaign types on broader ad platforms (e.g. Maximum performance/AI-assisted campaign types, Shopping and Search), your creative can be tailored to fit the AI experience and user intent of the moment.
This is why it is essential to remember that AI is a fluid and dynamic investment. Rigid creative demands (including pinning) make it difficult for AI creations to fully meet the needs of the human engaging with the AI.
If your brand has constraints (specific language to use, terms that can’t be included, etc.), most ad platforms are testing ways for humans to add constraints to how creative fits on AI surfaces. That said, if your brand must truly rely on specific creative that must always serve, you may not be able to leverage AI surfaces to the same extent as less restrictive brands.
It is worth emphasizing that AI-assisted campaign types (like Maximum AI and Performance Max) often have the best chance of appearing on AI surfaces due to their creative flexibility, broader matching, and dynamic audience mapping. That said, standard search and purchase formats may also be eligible depending on experience, market and query intent. Some platforms may also include rich creative formats (such as media style units) when they meet relevance and policy requirements.
Before we get into how to understand metrics, it’s worth remembering that AI surfaces are more than just AI assistants like ChatGPT and Copilot. AI modules also have their place on the search engine results page (AI Insights, Reply card formats on Bingetc.).
When the AI suggests something, it is important that the human does not have the impression that the recommendation is purely motivated by sponsorship. This is why many experiments clearly separate quotes and other non-ad modules from paid placements, and why ad eligibility is generally subject to a high relevance bar. Additionally, structured business information (e.g.: accurate pricing, availability, shipping, returns, and customer service details) helps AI systems present more trustworthy options and provides trust signals that reassure users that they are engaging with a legitimate supplier.
How to Think About AI Placement Metrics
Many make the mistake of viewing AI as a mere discovery or purely “bottom of the funnel” performance. In practice, AI can significantly reduce consideration cycles; sometimes taking a user from discovery to conversion in less than 30 minutes. In Copilot’s internal analytics, these placements showed up to 25% higher relevance than comparable SERP placements for similar intent.
At the same time, AI placements impose an even stricter bar for advertising relevance than conventional SERPs. This may lead to questions about volume as well as whether AI investments represent a significant standalone investment opportunity.
AI placements merge the line between branded and performance media buying. That’s why it’s critical to get the word out about these metrics and why conventional return on ad spend (ROAS)/cost per action (CPA) goals might not be as useful for AI surfaces.
Some AI experiments allow advertisers to create audiences based on engagement signals from those placements. Others are structured more like awareness purchases (e.g., CPM-based inventory), where the primary goal may be exposure and consideration rather than an immediate transaction on the platform. If you evaluate these placements based solely on last click conversions, they will often appear weaker than they really are.
Relying on data-driven attribution (which has been the standard on Google for some time) allows you to get a more complete picture of how different engagements led the user to say yes to you.
Yet this remains a “performance marketing” mindset. To take full advantage of AI placements, you also need to incorporate brand sentiment, share of citations, and other awareness metrics.
That’s why it’s critical to leverage AI visibility and on-site behavior tools to understand how often AI systems look to you for answers and what users do after they land. Session replay and UX analysis tools (e.g. Microsoft Clarity, Hotjar, FullStory, etc.) can help you spot friction, intent mismatches, and content gaps that matter for traditional and AI-based traffic sources.
When reporting is limited or aggregated, focus on directional measurement: compare conversion quality (not just quantity), monitor assisted conversion paths in data-driven attributionAnd run structured tests (geographic breakdowns, time-based holdbacks, or budget input/output experiments) to estimate the incremental increase. Pair this with brand-aware signals like direct traffic, branded search demand, and “citation share” in AI responses to avoid undervaluing top and middle funnel impact.
Establish a budget for AI internships
Let’s return to the original question at the heart of this article: are AI internships worth it?
If you think a 194% higher conversion rate (based on internal Microsoft data) is worth the creative flexibility that AI placements often need, then it’s worth planning how to build a budget and operational time around them. If you know you can’t say yes to this flexibility, the added value won’t matter because rigid compliance requirements will limit where and when how creatives can adapt. That’s why most major ad platforms continue to offer options that adhere to strict creative and policy constraints.
As has been shared previously, major advertising platforms can provide access to AI surfaces through existing campaign types. In many accounts, pricing appears similar to comparable non-AI inventory once you normalize for intent and competition, although actual cost will vary depending on market, query class, and available supply.
AI-powered advertising platforms may price inventory differently, often reflecting a more limited supply and tighter user experience constraints on the number of ads that can appear. In practice, this means you may need a sufficient daily budget to exit the market. learning phase and generate a signal (clicks, engagement, conversions) before you can judge performance. Instead of relying on a universal CPC, build a testing budget based on typical costs for your category, your conversion rateand the minimum volume you need to make a decision.
The other part of budgeting is the time it takes to create/manage creation, targeting and results. AI creatives include options to let people know more information about your product/service before clicking on your site/allowing the agent to complete the transaction via AI.
Takeaways
Here are the most important things to remember about AI ads:
- They won’t always be served, and if they are, it’s because the platform believes with a high degree of confidence that your ad will be a net benefit to the human engaging with the AI.
- Privacy considerations mean that distributed metrics for AI surfaces are more complicated than conventional reporting.
- Different AI surfaces apply different inventory valuation to locations. The budget that works for one may be higher or lower than another.
AI investments have been part of the marketing mix for years; they just have more visibility now. Whether you’re accessing them through existing campaign types or looking into AI-driven purchases, the central question isn’t just “are they worth it.” In many cases, data allows for testing, especially when you evaluate beyond the last click and account for incrementality.
The bigger question is whether your brand can say yes to the creative flexibility that the AI era demands. Creatives locked into rigid formats or constrained to a single, static landing experience tend to perform less effectively on AI surfaces than creatives that are able to adapt from intent to message to solution as user needs evolve.
More resources:
Featured image: Paul Poetry/Search Engine Journal
