For most of digital advertising history, the click had one job: get the user somewhere else.
An ad caught your attention. You clicked it. A landing page opened. You read the offer, filled out a form, booked a call, bought something or left.
An enormous industry was built around improving that journey.
Advertisers optimized CTR. Performance teams tested landing pages. Affiliates built funnels. Lead-generation businesses experimented with forms, quizzes and pre-landers. Conversion specialists argued about button placement, page speed and whether asking for a phone number too early was killing the funnel.
Now AI is beginning to challenge something much more fundamental than where the button should go.
What if there is no landing page at all?
Key takeaways
- In September 2026, OpenAI started testing Sponsored Agents: after clicking an ad in ChatGPT, a user can talk directly to the advertiser's AI agent.
- A conversation can qualify a prospect better than a static form, and it reveals why the person is there, not just what they clicked.
- The metrics that matter move deeper into the funnel: cost per qualified conversation, per appointment, per approved application.
- Traffic can be described by expressed intent, not only by GEO, vertical and format. That makes information about traffic more valuable.
From landing page to conversation
In September, OpenAI introduced a new generation of advertising experiences that included Sponsored Agents, a format designed to let a user move from an advertisement directly into a conversation with an AI agent from the advertiser. For now, it is being tested with select advertisers in the United States.
That sounds like a new ad format. It could turn into something much bigger.
Because once the advertisement can start a conversation, answer questions, understand what the user wants and potentially guide that person toward a transaction, the traditional boundary between the ad and the sales funnel begins to disappear.
The old journey looked roughly like this:
Ad → Click → Landing Page → Form → Lead
The emerging version could look more like:
Ad → AI Conversation → Qualification → Conversion
That changes what a click is actually for.
People want answers, not websites
A person looking for car insurance doesn't necessarily want to visit an insurance website. They want to know how much insurance will cost.
Someone looking for a loan doesn't necessarily want to read a lender's landing page. They want to know whether they qualify, how much they can borrow and what the monthly payment might be.
A traveler doesn't necessarily want to navigate ten hotel pages. They want a hotel that fits their dates, budget and requirements.
Digital advertising has historically responded to those intentions by sending people to websites and asking them to find the answer themselves.
AI can potentially reverse that relationship.
Instead of sending the user to information, the information comes to the user.
OpenAI's advertising announcement describes Sponsored Agents as experiences where people can engage directly with an advertiser's agent after interacting with an ad. The company is also building tools around AI-generated advertising creative and connections to commerce and CRM systems, with HubSpot and Shopify as the first partners, bringing the advertising interaction closer to the systems where businesses already manage products and customers.
The important word here is not sponsored.
It's agent.
A conventional ad can show a message. An agent can respond.
What a conversation does that a form can't
That difference becomes much more interesting when you look at it through the lens of performance marketing.
Imagine someone sees an ad for health insurance.
Today, the advertiser might send that user to a landing page containing several paragraphs of copy, a comparison table and a form asking for their age, location, phone number and email address.
An AI agent could simply ask:
Agent: What kind of coverage are you looking for?
The user answers.
Agent: Individual or family?
User: Family.
Agent: Which country do you live in?
User: Spain.
Agent: How many people need coverage?
User: Three.
Within a few exchanges, the advertiser may know considerably more about that prospect than it would have learned from a landing-page visit.
More importantly, the user hasn't been forced to navigate a website to provide that information. The qualification process has happened inside the conversation.
This is where Sponsored Agents become particularly relevant to lead generation.
The traditional online lead is essentially a packet of information collected through a form. Name. Email. Phone number. Perhaps a few qualifying fields.
A conversation can produce something much richer.
It can reveal what the person actually wants, what constraints they have, how urgent the purchase is and whether they meet the advertiser's criteria before the lead is ever passed to a sales team.
In theory, that means the industry could move from generating more leads to generating better-qualified ones.
For verticals such as insurance, lending, real estate, education, travel or B2B services, the implications are obvious. These are businesses where a conversion often requires questions before a transaction can happen.
AI is extremely well suited to asking questions.
It also doesn't have to ask every user the same ones.
A traditional form is largely static. The business decides which fields everybody must complete.
A conversation is dynamic.
If one answer makes three other questions irrelevant, an agent can skip them. If an answer suggests that the prospect needs a different product, the conversation can change direction. If somebody doesn't understand the offer, the agent can explain it immediately instead of hoping the landing page contains the right paragraph.
This doesn't mean websites are about to disappear. There are plenty of purchases where a page is still the best interface. People want to browse products, compare images, inspect specifications and sometimes simply read without having a conversation with anything.
But landing pages have never existed because consumers love landing pages. They exist because websites were the most practical interface available for turning advertising attention into action.
AI gives advertisers another interface.
New metrics for paid traffic
And once that happens, the economics of paid traffic start to look different too.
Performance marketers have traditionally spent enormous effort optimizing what happens after the click. Traffic acquisition and conversion optimization are often treated as separate disciplines: one team buys the user, another team tries to make the landing page convert.
Conversational advertising compresses those two layers.
The ad generates the interaction. The interaction qualifies the user. The qualification can potentially lead directly to the next commercial action.
That could also change how advertisers evaluate traffic.
Consider two campaigns generating 1,000 clicks each.
The first sends users to a landing page and produces 80 form submissions.
The second generates only 600 conversations but discovers that 150 users meet the advertiser's actual qualification criteria.
Which campaign performed better?
CTR alone won't tell you. Neither will the number of landing-page visits. Even a conventional cost-per-lead metric may become less useful if one system knows substantially more about the commercial quality of each prospect.
The important metric starts moving deeper into the funnel:
- cost per qualified conversation;
- cost per qualified prospect;
- cost per appointment;
- cost per approved application;
- cost per customer;
- revenue per conversation.
That shift is already familiar to sophisticated performance marketers. The industry has spent years trying to optimize away from superficial events and toward actual business outcomes. The same logic is behind Google moving Search campaigns toward AI Max, where the advertiser increasingly defines the result instead of the keyword.
Conversational AI could accelerate it.
First-party data and expressed intent
There is another reason this matters: first-party data.
A landing page tells an advertiser quite a lot about behavior. Someone arrived from a particular campaign, spent a certain amount of time on the page, clicked something and perhaps completed a form.
A conversation can tell the advertiser why that person is there.
That distinction is powerful.
If a user tells an insurance agent, "I'm moving to Spain next month with my wife and two children and need private health insurance," that is not simply another click from an insurance keyword.
It is expressed intent.
For advertisers, expressed intent can be extraordinarily valuable.
It also creates obvious responsibilities around privacy, consent and the handling of conversational data. The more information an AI agent collects during a commercial interaction, the more important it becomes to be clear about what is being collected, how it is used and when that information is passed into CRM or other business systems.
Those questions will become increasingly important if conversational advertising scales.
Attribution gets messier
So will attribution.
The traditional funnel is relatively easy to visualize because each stage has a familiar event attached to it: impression, click, landing-page view, form submission, sale.
A conversational funnel is messier.
When does a conversation become a lead?
- After the first message?
- After the user provides contact details?
- After the agent determines that the person meets the advertiser's requirements?
- After an appointment is booked?
The answer will probably vary by advertiser, which means the performance industry may need a new layer of measurement around conversational intent and qualification.
What it means for affiliates
Affiliates will have to think about this too.
A large part of affiliate marketing is built around controlling the journey between traffic source and advertiser. Publishers create content, comparison experiences, quizzes, pre-landers and other mechanisms designed to turn relatively broad traffic into commercially useful intent.
AI agents can perform part of that work dynamically.
A finance publisher, for example, may eventually send traffic into a conversational experience that determines whether the user is looking for a mortgage, personal loan, refinancing or another financial product before deciding which advertiser or offer is relevant.
At that point, the agent isn't simply replacing a landing page.
It starts behaving like an intelligent traffic router.
Traffic described by intent, not only by source
And that may be where things become especially interesting for traffic marketplaces.
The advertising market has traditionally classified supply using fairly broad labels: finance traffic, casino traffic, dating traffic, Tier 1 traffic, mobile traffic, search traffic, social traffic.
Those labels tell buyers where users came from.
Conversational systems can potentially tell them what users actually want.
That creates a much richer definition of traffic.
A publisher isn't simply offering 10,000 finance visitors. It may eventually be able to offer a pool of users who have expressed a specific need, answered qualifying questions and reached a defined level of purchase intent.
For buyers, that is a very different product.
And for marketplaces such as Traffberry, it points toward an interesting evolution. Advertisers will still need traffic. Publishers will still need buyers. But the information surrounding that traffic can become far more useful. As we wrote about AI agents becoming media buyers, the next generation of marketplaces will compete on how well they describe and qualify supply.
Instead of discovering supply only by GEO, vertical, format or volume, buyers may increasingly care about how traffic was qualified, what intent signals are available and how close the audience is to an actual commercial decision.
AI doesn't remove the traffic market. It potentially gives the market better information about what the traffic is worth.
Reasons to be cautious
There are still plenty of reasons to be cautious. Consumers may not always want to talk to a branded AI agent. Poorly designed agents could easily create frustrating experiences. Advertisers will have to decide how much autonomy agents should have, particularly in regulated sectors. And a conversational interface that feels helpful in travel could feel intrusive in finance or healthcare.
But the direction is worth paying attention to.
For years, digital advertising has focused on making the journey after the click shorter.
AI may take that logic to its conclusion:
- Instead of improving the landing page, remove it from the journey when it isn't necessary.
- Instead of asking every visitor to complete the same form, have a conversation.
- Instead of counting every lead equally, understand what the person actually wants before deciding what that lead is worth.
The click used to be the bridge between advertising and the conversion funnel.
Increasingly, the ad itself may become the beginning of the funnel.
For advertisers looking for new audiences, Traffberry provides a marketplace to discover media opportunities across publishers, GEOs and verticals. Publishers can list the traffic they have available and make it visible to buyers looking for new sources of acquisition.
The interfaces will keep changing. Landing pages may become conversations. Forms may become agents. Traffic buying itself may become increasingly automated.
But somebody still has to bring the right audience into that conversation.
And that is where the value of traffic begins.
Sources
- OpenAI: Reimagining advertising with AI (September 16, 2026)
- OpenAI Help: Set up ChatGPT Ads in HubSpot
- OpenAI Help: Set up ChatGPT Ads for Shopify