OpenAI Tests Agent Campaign Type That Replaces Landing Pages With AI Chat

OpenAI is testing a new paid ad format that skips the landing page entirely. According to ChatGPT Ads Business Agent Conversation Campaign Type, a new campaign type called Agent has appeared inside the ChatGPT Ads Manager for select advertisers — and instead of sending a click to your website, it drops the user directly into a configured AI business agent conversation.

The news was first spotted by Juozas Kaziukėnas, who shared details and a screenshot on LinkedIn. Barry Schwartz at Search Engine Roundtable confirmed he does not yet see the option in his own account, but noted Kaziukėnas described it as "in place in the ads manager for select advertisers." This is early and limited — but the implications run wide.

The Details: How the Agent Campaign Type Actually Works

This isn't just a chatbot widget bolted onto a display ad. According to Kaziukėnas's description as reported by Search Engine Roundtable, there are three distinct layers to how an Agent campaign is built and delivered.

First, OpenAI scrapes the advertiser's website to build a business profile. That profile captures common customer questions, support information, and general business context — essentially giving the agent a working knowledge base before a single conversation happens.

Second, the advertiser creates a business agent that follows defined instructions and draws on that profile for context. Critically, the agent also has access to product feeds and MCP tools for data access, and can capture leads through custom forms. That last part matters a lot — this isn't just a Q&A bot. It can act.

Third, the advertiser launches a campaign that points to that agent rather than a website URL. The underlying technology, according to the report, is the same as custom GPTs. So if you've built or used a custom GPT, you already have a rough mental model for what this experience looks like on the user side.

For context on what's currently available, the campaign type options Schwartz sees in his own ads manager are Standard and Product Feeds. Agent is not yet visible to everyone — this is a limited rollout.

Why This Matters for SEO and Paid Search Strategy

Picture this: you've spent months tuning a landing page — headline tests, form placement, load speed, the works. Now imagine a competitor's ad drops the same prospect into a conversation that knows their product catalog, can answer objections in real time, and captures a lead without the user ever touching a URL. That's the gap this campaign type is designed to close.

The shift here isn't just tactical. It's architectural. Traditional paid search is built around the click-to-page model. An Agent campaign breaks that model at the foundation. Conversion rate optimization, A/B testing frameworks, bounce rate analysis — all of that assumes a page is involved. Remove the page and you remove the entire playbook most performance marketers have been running for twenty years.

For SEO specifically, this raises a sharp question: if users increasingly resolve their intent inside a chat interface rather than a website, what happens to organic traffic as a conversion channel? I've been watching AI Overviews push zero-click behavior further up the funnel for a while now. Agent ads could do the same thing to paid traffic. The website as the primary conversion surface is under real pressure from multiple directions at once.

And yes, this isn't OpenAI breaking entirely new ground. As Schwartz notes in the piece, Google has announced business agents in its own ad products before. But ChatGPT's ad platform reaching this feature matters because of where ChatGPT users already are mentally — they came to have a conversation. An agent ad meets them exactly there.

What to Do Now

You don't need access to the Agent campaign type today to start preparing for it. Here's where to focus.

1. Audit your website content for agent-readiness. OpenAI scrapes your site to build the business profile that powers the agent. If your product pages, FAQ content, and service descriptions are thin, vague, or buried under JavaScript, the agent starts with bad data. Clean, crawlable, well-structured content wins here — same as it does in organic SEO.
2. Map your top customer questions. The agent is designed to handle common questions and support scenarios. Build a prioritized list of the 10 to 20 questions your sales and support teams answer most often. That list is the seed of your agent's knowledge base, whether you're using this format or building any other AI-assisted sales tool.
3. Review your product feed health. The Agent campaign type has access to product feeds for data. If your feed has stale prices, missing descriptions, or broken image URLs, the agent surfaces that garbage directly in conversation. Run a feed audit now.
4. Watch your ChatGPT referral traffic. If you're not already tracking traffic from ChatGPT in your analytics setup, set up source tagging that captures it. As OpenAI's ad platform scales, that data gets valuable fast. AI visibility tracking can help you monitor how often your brand and content are surfacing inside AI platforms.
5. Stay close to early adopter reports. This is available to select advertisers only right now. Follow Juozas Kaziukėnas and Barry Schwartz on LinkedIn for first-hand screenshots as the rollout expands. Don't wait for a case study — read the early signals.

Background and Context: Where This Fits

ChatGPT's advertising platform is still young. The company began rolling out ads inside ChatGPT search in 2025, starting with standard formats before expanding to product feed campaigns. The Agent campaign type, if it launches broadly, would represent the most significant structural departure from traditional search advertising the platform has made.

The broader arc here is one the search marketing conference circuit has been chewing on for the past two years: AI models are becoming the interface layer between users and businesses. First it was AI Overviews absorbing informational queries. Then AI-native search like Perplexity pulling transactional research. Now we're looking at AI models potentially absorbing the conversion moment itself.

The businesses that will win in this environment are the ones that treat their content and data as infrastructure — not just for ranking, but for powering every AI surface their brand might appear on, paid or organic. Agent ads are one more reason to get that foundation right.

If you want a fast read on how visible your brand currently is across AI platforms before this paid layer adds more complexity, Aergos has a free AI visibility checker that gives you a baseline in minutes.