AI Advertising Platforms: The Two Kinds, Who Sells What, and How to Choose

"AI advertising platform" means two different products, and most pages that rank for the phrase only cover one. The first is an ad platform you already use, with AI running the buying. The second is a platform that places your ads inside AI conversations, which is new inventory. This guide separates them, lists who sells what as of October 13, 2026, and gives you the questions to ask before you spend.

We should say where we stand: Thrad is the second kind. We are a demand-side platform for sponsored placements inside AI conversations, so read the evaluation section with that in mind. The facts about every other platform come from their own documentation.

What is an AI advertising platform?

An AI advertising platform is either an existing ad platform that uses machine learning to automate bidding, targeting and creative, or a platform that sells ad placements inside AI assistants and AI search. The two solve different problems:

Kind one: AI inside the ad platforms you already use

Google describes Performance Max as "a goal-based campaign type that allows you to access all Google Ads inventory," across Search, YouTube, Display, Discover, Gmail and Maps, where "Google AI mixes and matches these assets to render tailored ads across all inventory." Meta's Advantage+ campaigns and its Ads Manager assistant, covered in our Meta AI ads explainer, do the same for Facebook and Instagram. Every large ad platform now ships some version of this.

These products change how you buy, not where your ads appear. They are worth using, and they are not what most people mean when they ask about advertising in AI. If your question is "how do I reach people while they are talking to an AI assistant," kind one does not answer it.

Kind two: platforms that place ads inside AI conversations

This is the new inventory. The placement appears while someone is asking an assistant for help, and the targeting signal is the conversation itself. Four sellers document a live product today:

• OpenAI Ads Manager sells ads that "can appear below the end of a response" in ChatGPT for users on the Free and Go plans, billed on impressions or valid clicks through "a relevance-weighted, second-price auction," with a recommended starting CPC bid of $3 to $5. Ads are not shown near sensitive or regulated topics, and reporting during the test is aggregated views and clicks. Our ChatGPT advertising guide covers setup and budgets.
• Microsoft Advertising serves eligible ads from existing Search, Shopping and Performance Max campaigns inside Copilot answers. There is no separate Copilot campaign, advertisers cannot opt out, and there are no Copilot-specific metrics.
• Amazon Ads places Sponsored Products and Sponsored Brands in Rufus shopping prompts, auto-enrolled from existing campaigns on a cost-per-click basis, in the US.
• Thrad is a DSP for ads inside AI conversations: advertisers bid on conversational intent in real time, with delivery in under 500 milliseconds, across 20+ countries, with one budget and one report across every AI app. The inventory is independent AI publishers, vertical assistants and specialized LLM tools rather than one walled garden, and the self-serve platform has no minimum spend.

Three of the most-asked-about assistants do not sell placements. Anthropic says Claude will remain ad-free. Google's ads leadership has said there are no ads in the Gemini app and no current plans to change that, while ads do appear in AI Overviews and are being tested in AI Mode through existing Search campaigns. Perplexity tested ads in 2024 and phased them out. Our comparison of which AI assistants sell ads tracks each one.

How do you evaluate an AI advertising platform?

Whichever kind you are buying, six questions separate a platform from a press release:

• Reach: which AI apps or surfaces, in which countries? One assistant is one audience; the AI open web is many.
• Signal: does targeting use the live conversation, or audiences and keywords carried over from search? Only the first is new.
• Placement: inside the answer, below it, or alongside? Ask to see a real example in a real chat interface.
• Controls: which categories and contexts are excluded, and can you add your own exclusions?
• Price and auction: CPM or CPC, first- or second-price, and what the minimum is. OpenAI publishes its auction type; ask every other seller for theirs.
• Reporting: aggregate views and clicks, or placement-level data you can join to your own analytics? Copilot reports nothing Copilot-specific today; plan around that.

Which kind should you buy?

• If you want better results from Search, social and video budgets you already run, use the AI inside those platforms. It is included in what you pay.
• If you want to be present when people ask an assistant for a recommendation, you need kind two, and you need to buy it from more than one seller to cover more than one assistant.
• If you want one line on the media plan rather than four logins, Thrad lets you run one budget across the AI apps your customers already use. Launch a campaign from self-serve, or read how programmatic buying for the AI era works.

A checklist for choosing an AI advertising platform

1. Write down which meaning you need: automation on existing channels, or placements inside AI conversations.
2. List the assistants your customers use, by country, and check which of them sell placements at all.
3. Ask each seller for a live example of the placement, the auction type, the exclusions and a sample report.
4. Start with a self-serve test on a fixed budget before signing a managed agreement.
5. Measure AI placements as their own channel, with their own tracking, before folding them into always-on plans.