For 25 years, the website sat at the center of the buyer journey. Search, click, browse, convert. That sequence is compressing. Consumers now do their research inside ChatGPT, Claude, and AI Overviews and they reach the brand late in the process, or not at all.

The visits that still happen carry more intent than marketers have measured before. US ecommerce AI referral traffic from ChatGPT grew 210.5% year over year in Q1 2026, according to EMARKETER’s Industry KPI data provided by Grips Intelligence. Those visitors are also proving valuable: ChatGPT referral traffic converted 31% higher than non-branded organic traffic in 2025, per Visibility Labs data cited by Search Engine Land.

The same pattern shows up in conversations rather than clicks. Invoca’s analysis of 70 million voice and SMS conversations found that calls originating from ChatGPT carry a 49% lead rate, the highest of any channel we measure. Traffic is down and qualification is up.

A 2026 experience, then a 2016 one

That shift is clearest in high-consideration categories. A buyer might use a large language model (LLM) to compare options and narrow the list, then visit a single brand ready to decide.

Some brands have started building for that buyer. One national real estate platform lets people describe the home they want in plain language, such as three bedrooms, a yard, under $500K, and in this neighborhood. The platform then brings up listings with a written explanation of what fits and what doesn’t. It brings conversational search on the brand’s own property rather than ChatGPT’s. An agent doing the work a filter menu used to do.

Most brands do the opposite. The buyer gets a 2026 experience inside the LLM and a 2016 experience on arrival: static content, a filter menu, and a form promising someone will follow up.

That gap does more than disappoint. It teaches the best-qualified buyers to stay on the AI platform, and it compounds every quarter it goes unaddressed. Invoca research found that only 36% of buyers receive a response within an hour, and say 79% they’ll move to a competitor that responds faster.

The harder work starts after the citation

Most marketing teams are racing to get cited in AI answers. Fewer than 10% of sources cited in ChatGPT, Gemini, and Copilot rank in Google’s top 10, suggesting that the old playbook alone won’t be enough to earn AI visibility. But visibility is only the first step, and it already has marketers’ attention. The harder question is what happens after the citation works.

The brands with an answer are doing two things. First, they connect ad spend to where revenue actually lands, across ecommerce, the contact center, and retail, allowing them to evaluate new AI ad surfaces based on business outcomes. Second, they rebuild the owned experience around AI agents that can respond when intent appears, whether the buyer chooses a form, a text, a chat, or a call.

If ChatGPT and your brand experience had a love child, what would it look like? Most CMOs cannot picture it yet. That is the work.

The question worth answering before the next budget cycle is what a brand experience becomes when the website is no longer the center of the customer journey. The funnel is not disappearing. It is being rebuilt around conversations, and the marketers who lead that rebuild will have more influence over revenue than the homepage ever gave them.