The 60% zero-click figure is the headline. It hides three very different numbers underneath.
Desktop searches end without a click 60% of the time. Mobile searches end without a click 77% of the time. Searches conducted in Google's new AI Mode end without a click 93% of the time.
That spread matters. Because the device your prospect is using when they search for your services now determines, to a significant extent, whether they ever see your website — regardless of how well you rank.
Mobile screens are smaller. The features Google adds above organic results take up disproportionately more vertical space. On a desktop, an AI Overview occupies maybe 25% of the visible page. On a mobile screen, the same Overview can fill 80%+ of what the user sees before scrolling.
The result: by the time a mobile user has finished reading the AI Overview, scrolled past the People Also Ask box, glanced at the Map Pack, and arrived at the first organic listing, they've often already made their decision — or been distracted by something else, or run out of time.
Mobile traffic also skews toward queries that don't require deep website exploration. "Plumber near me", "industrial electricians [town]", "skip hire same day". For these, the user wants a phone number, a price, or an availability check — all of which the SERP now delivers directly. The visit to your website would be redundant.
Three structural factors push mobile zero-click higher:
Smaller viewport. More of the page-above-organic per scroll. Less reason to scroll further.
Tap-to-call buttons in the SERP. Particularly on Map Pack results. The user calls without ever visiting a site.
Tighter intent windows. Mobile searches are often "I need this now" queries that don't require website research.
For a construction firm, the implication is direct: a significant share of your mobile search visibility now happens in places that aren't your website — primarily the Map Pack and AI Overview. If you're not present in those, you're not present.
AI Mode is Google's conversational search interface, rolled out to logged-in users through 2025 and 2026. It looks more like a chat with ChatGPT than a list of links.
The user types a question. The AI responds with a synthesised answer, citing two to four sources. The user follows up with refining questions. The conversation continues until the user has what they need.
There is no "list of ten blue links" in AI Mode. There are citations — small numbered references the user can click to verify a claim — but the dominant interaction is reading the AI's answer, not visiting any particular site.
The 93% zero-click figure makes sense in this context. Users are in AI Mode precisely because they want a synthesised answer, not a list of websites. The clicks that do happen are mostly verification clicks ("let me check that"), not exploration clicks ("let me learn more").
For construction queries in AI Mode, the dynamic is:
If your firm is one of the named firms, you're in the consideration set. If not, you don't exist for that user in that session.
There is no organic ranking that fixes this. There is no SEO trick that overrides which firms the AI chose to name. The only way to be named is to be sufficiently present in the underlying training and reference data that the AI considers your firm relevant.
If your firm's traffic split is something like 65% mobile / 30% desktop / 5% AI Mode in 2026, your invisible-to-the-user percentage is closer to 70% than to 60%. Most of your search "visibility" produces no website visit.
That doesn't mean SEO investment is wasted. It means the investment needs to be aimed at a different target than "rank for these keywords".
Three reallocations worth considering:
From keyword targeting to citation targeting. Less time on optimising title tags and meta descriptions for specific keywords. More time on producing content structured to be quoted inside an AI Overview — direct answers, comparison tables, owner-attributed quotes, specific data points.
From organic to local pack defence. For trades where your buyer is local, the Map Pack is now your most resilient piece of search real estate. Google Business Profile completeness, review velocity, consistent NAP, and category-specific optimisation matter more than they did five years ago.
From SEO-only to LinkedIn + SEO. Branded searches (people Googling your firm by name) bypass the entire above-the-fold dynamic that's eating organic traffic. The fastest way to grow branded searches for a B2B construction firm in 2026 is sustained, owner-led LinkedIn presence — not company-page content, not paid ads. Real activity from the founder or MD.
A few specific tactics retain their value despite the device-driven decline:
Click-to-call from Google Business Profile. Direct calls from the Map Pack convert at high rates and bypass the website entirely. If your GBP isn't fully optimised — categories, services list, attributes, hours, response time — that's a higher-leverage fix than most organic SEO work.
Site speed for the visitors you do get. Mobile users who do click through are now further down the funnel and lower-tolerance for slow pages. Core Web Vitals matter more than they did in 2020 — a 3-second mobile load now loses you ~30% of the comparison-shopping clicks you're still getting.
Schema and structured data. AI Overviews preferentially cite content with clear schema markup. Construction-specific schema (LocalBusiness, Service, FAQ, HowTo) increases the probability that your content gets pulled into the answer rather than someone else's.
Owner-attributed content. Quotes, opinions, and case study commentary attributed to a named owner or director get cited at higher rates than anonymous corporate content. AI tools weight authored content above brand content.
If you'd like to know what your firm currently looks like in mobile SERP, AI Mode, and across the major AI tools, Quiet Giants runs an audit specifically for £2–30m construction firms. One firm per sector. No retainer pitch. Book the call.
Sources: SparkToro Zero-Click Study 2024; Similarweb Mobile Search Report 2026; Semrush AI Mode Analysis 2026.