The customer journey moved.
Before AI, the full journey lived on your website. Users browsed across sessions, compared options, and converted. The funnel was visible, measurable, designable. That's changed. Not slowly. Here's what the data shows.
The before: what the traditional funnel looked like
In a standard funnel, thousands of visitors produced a small percentage of conversions. The journey was long, inefficient, visible.
But the entire funnel was yours to design, measure, and optimise.
192 - 272 days
average B2B buying journey
27
average touchpoints before close
75%
B2B bounce rate, pre-AI baseline
0.6%
average B2B sitewide conversion rate
Pre-AI baseline, 2015-2022. Sources: Contentsquare, Forrester, Gartner, Baymard Institute, Dreamdata, Adobe Analytics
What AI changed: the journey moved off your site
AI did not remove the customer journey. It redistributed it. Discovery, comparison, and evaluation now happen inside AI interfaces before a user ever reaches your website. By the time someone clicks through, they have often already decided.
Your website is no longer where the journey happens. It is where it ends.
84%
of B2B buyers say AI speeds their research process
77%
of B2B journeys now complete in 12 weeks or less
Sources: SparkToro / Datos zero-click search study, 2024; Bain & Company, 2025;
Google + National Research Group B2B buyer study, 2025
Is your conversion rate up for the right reasons?
A widely repeated claim is that AI increases conversion rates by 20 to 30%. The number is not always wrong. But it is almost always misunderstood.
Conversion rate is conversions divided by traffic. AI filters out low-intent visitors before they click, shrinking the denominator. Higher conversion rate — not from more conversions, but from fewer unqualified visitors.
The real question is how much demand is resolving inside AI before the click, versus how much is being sent your way.
+31%
conversion rate for AI-referred visitors vs non-AI
+45%
time on site for AI-referred visitors
20 - 50%
potential decline in organic traffic overall
Sources: Adobe Analytics, holiday season data, 2025; McKinsey & Company digital traffic analysis, 2025
The sites AI cites are not the biggest. They are the best structured.
AI does not browse. It extracts. It pulls specific, clearly written, well-structured content from pages that make it easy. Sites that are invisible to AI are not necessarily bad sites. They are just not built for how AI reads.
That is a fixable structural problem, not a budget problem. It sits in the UX and content layer, above the technical foundations most tools audit. We work across a range of industries.
What does AI compression mean for your numbers?
The impact is not uniform. It depends on your intent split, your category's sensitivity to zero-click search, and how well your content is structured for AI discovery.
Run the simulator below to model how these shifts affect your traffic, conversion rate, and total volume.