Your Website Funnel Just Got Shorter. Rebuild For It
Jillian Oco, CMO
Fewer visitors are reaching MSP websites at all, and the ones who do are arriving later in their buying decision than they used to. That changes what website architecture is actually for. Instead of building pages to attract and nurture unfamiliar traffic, the job now is capturing a smaller number of already-informed visitors before they leave without converting.
How much smaller is the top of the funnel getting?
The shift is bigger than most marketing teams have priced in. SparkToro's analysis of Similarweb clickstream data found that 68.01 percent of Google searches in the United States ended without a click during the first four months of 2026, up from 60.45 percent just two years earlier, with AI Overviews appearing on more than 20 percent of searches and cutting click-through rates on those queries by nearly 60 percent.
That is not a gradual SEO trend an MSP can out-optimize with more blog posts. It is a structural reduction in how many people ever reach a website's front door, which means every page that does get a visit needs to work harder than it used to.
Most website architecture in the channel was built for the opposite assumption, that traffic would keep growing if the content kept improving. Homepages, service pages, and blog posts were stacked in a funnel shape meant to catch a large number of unfamiliar visitors and slowly walk them toward a form fill. That shape made sense when most visits came from someone typing a question into Google and clicking through several results to compare answers. It makes less sense when a meaningful share of that comparison now happens before a browser tab ever opens.
Who is actually showing up, and what do they already know?
The visitors who do arrive are frequently better informed before they land. G2's 2026 AI Search Insight Report, surveying 1,076 B2B software buyers, found that 51 percent now start their research with an AI chatbot more often than with Google, up from 29 percent less than a year earlier, and that 71 percent rely on AI chatbots at some point during their research process.
A buyer who has already asked ChatGPT or Gemini to compare managed-service options has typically formed an initial shortlist before ever opening a browser tab to visit a site directly. That buyer does not need an explainer on what an MSP does. They need confirmation that this specific provider matches what the chatbot already told them, and a fast way to act on it.
Is this traffic actually worth building for?
It is becoming measurably real, even if it is still a small share of total traffic today. Demandbase's analysis found ChatGPT-referred visits to B2B websites grew 303 percent year over year, from roughly 645,000 monthly visits in June 2025 to 2.6 million in June 2026. The absolute numbers are still small next to organic search, but the growth rate means this referral path will keep mattering more with each passing quarter, not less.
Google's own guidance adds a useful detail about who clicks through at all. Google states that clicks arriving from search results pages with AI Overviews are higher quality, meaning those visitors are more likely to spend more time on the site. Fewer visits, but more engaged ones, is the pattern showing up across both AI chatbot referrals and AI-influenced search clicks, and it points to the same architectural conclusion either way.
What does this mean for how a site should be built?
A homepage designed to walk a cold visitor through "what we do, why it matters, who we've helped" before ever mentioning a next step is built for a visitor who mostly does not exist anymore in the same numbers. The visitor who does land on that page today is more likely arriving already knowing roughly what an MSP does and is checking specific details: pricing structure, service scope, and whether this provider looks credible enough to contact.
That argues for moving the conversion ask higher on every page that is likely to receive this kind of visit, not just the homepage. A product or services page buried three clicks deep, reachable only through a nav menu a chatbot-referred visitor has no reason to explore, is effectively invisible to this traffic even when the content itself is strong.
Where should the conversion ask actually live now?
Every page capable of answering a specific buyer question needs its own direct path to contact or a quote, not a single contact page the whole site funnels toward. If a visitor lands on a page about managed backup services because that is what an AI assistant pointed them to, the ask to talk to someone should be visible on that exact page, not three navigation clicks away on a generic "contact us" page built for a different kind of visitor.
| Visitor type | What they already know | What the page needs to do first |
|---|---|---|
| Cold organic visitor | Little to nothing about the provider | Explain the offering and build initial trust |
| AI chatbot referral | Has likely compared options already | Confirm fit and surface a fast next step |
| AI Overview click-through | Arrived with higher intent, per Google's own data | Answer the specific question, then convert |
Consider a hypothetical MSP whose services pages are built like brochures, with the only call to action sitting in a sticky header link. A visitor arriving from an AI assistant already past the "what is this" stage has to scroll past paragraphs written for someone earlier in the funnel before finding any way to act, and plenty of that traffic simply leaves instead of hunting for it.
Rebuilding architecture around fewer, more qualified visits
This does not mean abandoning top-of-funnel content entirely. Cold organic search and referral traffic from partners still exist, and plenty of visitors are still early in their thinking. It means every page needs to serve both audiences at once, giving enough context for someone unfamiliar while putting the conversion path early enough that someone who already knows what they need does not have to dig for it.
In practice that usually means moving a contact or quote link out of a sticky header alone and placing it directly after the section that answers the core question a page covers, rather than waiting until the bottom. It also means auditing navigation depth. A page that only a menu can reach, with no direct link from anywhere an AI assistant or search result might point to, is effectively unreachable by the exact traffic this shift is creating.
Structured data and clean page hierarchy matter here for a different reason than ranking. A page that clearly states what it covers, who it serves, and what happens next is easier for both a human skimming on a phone and an AI system summarizing it to a buyer to parse correctly. Getting that structure right across an entire site, not just a handful of flagship pages, is exactly the kind of audit ActiScore is built to run, scoring where a site's architecture is still built for a funnel shape that traffic patterns have already moved past.
What this is worth fixing first
Start with the pages most likely to receive specific, already-informed traffic rather than the homepage. Service and product pages, comparison content, and anything likely to come up when someone asks an AI assistant to shortlist providers deserve a visible next step above the fold, not just a mention further down the page. Running that audit manually across dozens of pages is slow, which is where tools built for this exact problem earn their keep, whether that means comparing options through the stack builder or reviewing the broader catalog of white-labeled tools built for this shift.
The funnel has not disappeared. It has gotten shorter and arrives later, which means the conversion ask has to move earlier on every page capable of catching it. See the full stack to see what auditing a site for that shift actually looks like in practice.