When will marketing catch up to how consumers actually use AI?

September 24, 2026

AI experience with fewer users could generate highly engaged interactions in which consumers spend significant time actively explaining their needs. Image credit: AI rendition in Shutterstock

By Bishesh Khadka

AI has given contextual advertising a significant upgrade.

Marketers can now understand meaning and relevance at a level that goes well beyond matching a keyword to a page. But while the technology for understanding consumers has advanced rapidly, much of the infrastructure used to monetize that understanding still operates according to an older assumption: advertising is fundamentally about buying and selling inventory.

Digital advertising was built around standardized units – pages, placements, impressions, search queries and feeds. A publisher had inventory. An advertiser wanted attention. Programmatic systems connected the two efficiently and at enormous scale.

Conversational AI challenges that model because consumers increasingly aren't simply navigating pages or submitting isolated queries. They're having conversations.

A conversation reveals something much more valuable than where someone happens to be online: what that person is trying to accomplish.

From target keywords to intent
Search provides marketers a powerful indicator of consumer intent.

If someone searches for "best running shoes," marketers can reasonably infer what that person wants.

But consider how the same interaction unfolds with an AI assistant.

A consumer might begin by asking how to start running, explain that she’s previously had trouble finding comfortable shoes, specifies the surfaces she runs on, narrows down a budget and eventually asks for recommendations.

No single keyword captures that journey.

Where traditional contextual advertising evaluates the content around a consumer at a particular moment, conversational systems can help illuminate a consumer’s immediate intent as it develops.

The intent becomes clearer as the conversation develops. That's not a new skin on keyword targeting. It's contextual advertising, but much more dynamic.

It’s important to note, however, that advertisers should never influence the assistant’s underlying organic response. Sponsored content should remain clearly disclosed and independently selected.

Bishesh Khadka

Infrastructure still thinks in inventory
The challenge is that the advertising ecosystem wasn't built to value this kind of signal.

Programmatic advertising became successful in part because it standardized media and made transactions extraordinarily fast.

DSPs, SSPs and exchanges can evaluate and clear opportunities in milliseconds. That speed remains essential.

But conversational environments introduce a different question.

Instead of simply asking, "What impression is available?" marketers increasingly must ask, "What does this consumer need right now, and what commercial message (if any) would add value?"

Answering this question requires interpreting semantic context and decision stage in real-time, then translating those signals into something existing advertising systems can use without adding unacceptable latency.

It also means reconsidering how you evaluate bid opportunities.

Volume has traditionally been an important measure of digital supply. Yet an AI experience with fewer users could generate highly engaged interactions in which consumers spend significant time actively explaining their needs.

Treating those interactions exactly like conventional pageviews risks undervaluing them.

Intent can span the entire funnel
Conversational intent shouldn't be understood solely as a bottom-of-funnel performance signal.

When a conversation is broad and exploratory, an awareness message may be appropriate.

As the consumer clarifies needs and moves toward consideration, more targeted advertising can become useful.

A consumer actively comparing products may benefit from product-specific information. After a purchase, the relevant commercial experience changes again.

In other words, advertising value doesn't suddenly appear at a single point in a conversation. It evolves alongside the consumer. That creates an opportunity to make advertising feel less like an interruption and more like useful discovery.

However, marketers should be careful not to equate richer context with permission for greater surveillance.

Interpreting semantic signals from immediate interactions does not require access to consumer identity or long-term browsing history, which is a critical advantage of conversational context.

It goes without saying that appropriate disclosure, consent and compliance with applicable privacy requirements still apply, as consumers might consider AI conversations more personal than ordinary web browsing.

An important distinction exists between understanding immediate context and building a persistent profile for a user.

Start with AI supply you already have
DSPs and agencies don't need to wait for the entire advertising ecosystem to be rebuilt before they begin learning.

Many organizations already manage digital properties that are adding AI experiences.

A standard website might now include a chatbot or conversational discovery tool. These interactions produce signals about what consumers want, yet they may sit entirely outside the site's existing monetization strategy.

This AI-enabled supply becomes a testing ground for teams looking to establish defined experiments and concrete success criteria for using AI. They can:

Evaluate semantic context in conjunction with conventional measures

Test how advertising relevance changes at different stages of a conversation

Determine which signals to incorporate into existing buying infrastructure while maintaining the speed programmatic advertising requires

The goal isn't to discard the advertising stack that already works. It's to expand what that stack can understand.

The interfaces consumers use will continue changing.

Currently, the dominant experience may be a one-to-one chatbot. Tomorrow, it could involve multiple AI agents, group conversations, dynamically generated images or real-time video.

The common denominator is intent.

These experiences exist because a consumer is asking a system to do, discover, create or decide something.

AS AI BECOMES embedded across the internet, intent won't become a rarer signal. It will become omnipresent.

Marketers that learn to recognize and responsibly act on that signal won't simply have better targeting. They'll be building an advertising model that finally reflects how consumers are beginning to use the internet.

Bishesh Khadka is founder/CEO of San Francisco-based Imprezia, which claims to be the world’s first AI native ad network.

Share your thoughts. Click here

When will marketing catch up to how consumers actually use AI?

When will marketing catch up to how consumers actually use AI?

Traditional agency model outmoded – and what should replace it

New challenge for brands: Cost of running AI at scale

68pc of small business owners to raise marketing spend this year

Why your TV budget needs a partner

5 common CRM mistakes to avoid making at all costs

Free-trade deals with US, EU turn India into superhighway for global luxury

Make your content AI-citable: New rules of structure, social and signals

How AI is transforming marketing measurement

5 trends that will define small-business marketing in 2026