The landing page is no longer the first conversation with a customer

For years, the pattern of online advertising was fairly simple.

A user saw an ad, clicked and arrived at a website. Only there would they learn the details of the offer, compare options, look for answers and decide whether to submit a form or make a purchase.

That model is not disappearing. But another layer is beginning to appear between the ad and the website:

a conversation with AI.

OpenAI is currently testing Sponsored Agents in ChatGPT, a format that lets someone start a conversation with an AI agent representing a business after clicking an ad. The user can ask about its products or services while still in ChatGPT. The format is currently available only to selected advertisers in a limited alpha.

Google is moving in a very similar direction. In its September Demand Gen updates, it announced Business Agent for YouTube Ads. With video ads using a product feed, users can ask questions about a brand or product without leaving their current YouTube viewing context. Google is currently inviting US merchants into beta testing.

That does not mean we are about to stop needing landing pages.

Something else is changing.

Customers may ask their first questions before visiting the website

Imagine a simple situation.

A business is running a campaign for a preschool. Today, someone sees the ad, visits the website and starts looking for answers:

• are places still available,
• what age children are accepted,
• how the settling-in process works,
• what hours the preschool operates,
• how much it costs,
• where the preschool is located.

In a conversational model, they could ask some of those questions before seeing the website at all.

The same applies to ecommerce.

Instead of entering a store, browsing five products and comparing descriptions, a customer could ask:

“Which of these products is suitable for dry skin?”

“Is this model compatible with my device?”

“What is the difference between versions A and B?”

“Is the product available in size XL?”

Only after receiving an answer would they move on.

And this is where the interesting change begins.

An ad is becoming more than a creative asset

Until now, much of the work has focused on:

visuals + ad copy + the landing page.

If an AI agent appears between the ad and the website, another part of a campaign becomes the quality of information a business makes available about its offer.

Because the agent needs to know:

• what the business actually sells,
• whom a product or service is for,
• how the variants differ,
• what the prices are,
• what terms apply,
• what the business does not offer,
• how delivery, booking or contact works,
• what customers ask most often.

If that information is scattered, outdated or contradictory, the problem no longer affects only a person browsing the website.

It can also affect an AI system trying to answer a potential customer based on that data.

The website is not becoming less important

In my view, quite the opposite.

A good website is beginning to serve two roles at once.

We have known the first for years:

it needs to persuade a person.

The second will become increasingly important:

it needs to describe the business, its offer and its products clearly to the systems that mediate between customers and brands.

That is why AI's development is not a reason to care less about the website.

It is a reason to stop treating it solely as an attractive business card.

If a website says a lot but, after reading it, you still cannot answer basic questions:

What exactly do you sell?
For whom?
How much does it cost?
How is one service different from another?
What should I do next?

then both people and AI will encounter the problem.

FAQs may become useful again — but not as a section made for SEO

For years, FAQs were often treated as an SEO element.

Businesses added a dozen or so questions, sometimes aimed more at search engines than at customers' actual needs.

In a conversational model, an FAQ takes on a much more practical role.

Well-prepared questions and answers can be an organised source of information about a business.

This is not about inventing fifty artificial questions.

It is better to answer ten things customers really ask, clearly:

Can I cancel?
How long does delivery take?
What happens at the first meeting?
What is included in the price?
Do you work with businesses throughout Poland?
Can the product be returned?
Which option should I choose?

This information is useful whether a person, Google or an AI agent reads it.

Ecommerce will need even better product data

Online stores have another layer to consider.

AI needs more than good marketing copy.

It needs specific data.

Material. Size. Variant. Ingredients. Colour. Use. Availability. Compatibility. Price. Technical specifications.

If that information exists only in an image or is entered inconsistently across a store, a system will find it harder to understand a product and compare it with others.

As AI develops, less glamorous things therefore become more important:

a well-prepared product feed, organised attributes, logical product names and specific descriptions.

Not because this is another “AI hack”.

Machines simply need structured data to understand the offer correctly.

What can a small business do now?

I do not see a need to create a separate, grand “AI strategy” just because Sponsored Agents or Business Agent have appeared.

Both are still in development and individual features have limited availability. OpenAI describes Sponsored Agents as a closed alpha, while Google is testing Business Agent in YouTube ads in the US.

There are, however, several things worth doing regardless of how quickly these products develop.

Organise the offer.

Each service should clearly describe its scope, audience and next step.

Check the questions customers ask most often.

If a salesperson answers the same question several times a week, the answer probably belongs on the website too.

Make information consistent.

The price on the website, in the store, in the feed and in ad materials should not be four different pieces of information.

Improve product data.

In ecommerce, the feed and product attributes should be treated as an important part of marketing, not merely a technical Merchant Center requirement.

Watch the measurement.

If a customer talks to AI first and only later visits the website, we will need good GA4, correctly configured conversions and information about where the enquiry really came from even more.

The landing page is not dying. Its place in the process is changing

I do not think we will stop building websites and landing pages any time soon.

We still need somewhere to:

• show the full offer,
• build credibility,
• present work examples,
• collect a form,
• make a booking,
• complete a purchase.

What changes is the moment when someone first encounters information.

Until recently, customers had to enter a website to learn more.

Now part of that conversation can happen:

in Google, YouTube, ChatGPT or another AI system.

That means business marketing will increasingly involve more than bringing a user to the website.

It will also mean taking care of what can be understood about the business before someone visits that website.

That is why an organised offer, useful FAQs, good product data and consistent information about the business may matter more in the coming years than another impressive slogan.

AI may take over part of the first conversation with a customer.

But for it to do that well, the business must first know exactly what it wants to tell them.