Most B2B leaders have already accepted the premise. Buyers click less.
When a Google AI Overview appears, Ahrefs found the top-ranking page now gets 58% fewer clicks, based on 300,000 keywords. A year earlier, the same study measured a 34.5% drop.
The harder problem is what that does to reporting. Many tracking and attribution models were built around the click, and so were the metrics they produce, from traffic to cost per lead. If the click is losing meaning, so is every report that starts with it.
The click was always a proxy
The click was never the outcome. Pipeline is the outcome. The click just stood in for three things at once:
- Discovery. A buyer found you while looking for help.
- Intent. They cared enough to leave the search results page.
- Attribution. You could see the visit and follow it toward a deal.
For two decades, those three signals arrived together in a single event. That made the click easy to report and easy to target.
AI answers pull them apart. A buyer can find your brand in an AI summary, form a view of your category, and add you to a shortlist without ever visiting your site. Discovery happened. Intent formed. Attribution disappeared.
The behavioral data points the same way. Pew Research Center tracked the browsing of 900 US adults across 68,879 Google searches in March 2025. When an AI summary appeared, people clicked a traditional result on 8% of visits, compared with 15% when there was none. They clicked a link inside the summary itself on just 1% of visits.
The size of the effect is still moving. Seer Interactive's 2026 study of 53 brands found organic click-through rate on AI Overview queries fell to 1.3% in December 2025, then recovered to 2.4% by February 2026. A baseline you set six months ago may already be wrong. That makes this a measurement problem, not just a traffic problem.
The shortlist forms where you can't see it
Buyers who click less are still researching. They're just doing it somewhere you can't watch.
Forrester found that buyers using AI instead of traditional search are one-tenth as likely to click through to a vendor's site, and that B2B companies have seen traffic declines of 10% to 40% over the past year as buyers adopt AI. Buyers now use AI to research product information (54%) and to compare offerings (55%).
Adoption figures depend on who's asked:
- Forrester reports that 94% of business buyers now use AI in their buying process, up from 89% a year earlier.
- TrustRadius, surveying 1,862 technology buyers in January 2026, found 63% used AI during their purchase journey. Of those, 94% fact-check its answers at least some of the time.
The gap reflects different samples and definitions, but both agree on direction: buyers who verify AI answers still visit sources, but they arrive to confirm, not to explore. TrustRadius also found that 83% of buyers shortlisted three or fewer products.
The finding that should change how you report comes from Bain & Company. Bain notes that research continues to show about 90% of buyers purchase from their Day 1 list: the vendors they have in mind when the buying process starts. Bain's sharper point is that traditional metrics like brand awareness, qualified leads, and pipeline coverage describe familiarity and activity. They don't predict who wins. Many teams over-invest in what's traceable and under-invest in the moments that decide whether they're considered at all.
Put those together. The most important decision in your pipeline, whether you make the list, increasingly happens before the first click and outside traditional attribution channels.
What you can still measure
You won't get the click's old certainty back. You can get a clearer picture than the click ever gave you. Five instruments, in order of effort:
Self-reported attribution. Add a required "How did you hear about us?" field to demo and contact forms, with a free-text option. Buyers will name ChatGPT, a peer, a review site, a podcast. It costs almost nothing, and most B2B teams still don't do it.
Sales-logged influence. Ask reps to note when a prospect references an AI answer, a review, or a peer recommendation, and what it said. Sales conversations surface influence analytics can't see.
AI referral traffic, on its own line. Visits from ChatGPT, Claude, Gemini, Perplexity, and similar assistants usually carry an identifiable referrer. Give them their own channel grouping in your analytics so they aren't buried in referral or direct. Expect small numbers. The SEO Works tracked 95 websites from September 2025 to August 2026: organic search sent 39 times more traffic than AI assistants, with AI reaching about 3.9% of organic volume by August 2026. Small doesn't mean unimportant. It does mean you shouldn't judge the channel on volume.
Presence and accuracy in AI answers. You can't see what buyers ask, but you can ask the same things. Pick a fixed set of prompts that mirror how your buyers research: the problem, the category, vendor comparisons. Run them monthly across the major assistants. Record whether you appear, how you're described, and who appears alongside you. Accuracy matters as much as presence. A buyer who arrives with the wrong idea of what you do becomes a stalled deal later. In Seer's 2026 data, brands cited in an AI Overview earned about 120% more organic clicks per impression than uncited brands, though Seer is clear that's a correlation, not proof of cause.
Branded search and direct traffic, with care. It's tempting to treat rising branded demand as proof that AI visibility is working. The evidence doesn't support that yet. The SEO Works tested it directly and found no link between AI referral growth and direct traffic growth once overall site growth was accounted for. Track branded demand as context. Don't credit it to any one channel.
A measurement framework leadership will act on
A new framework fails if it stays inside marketing. The goal is reporting leadership trusts enough to make budget decisions from, even though it's less precise than the dashboard it replaces. Organize it in three tiers, each with its own job.
Tier one: Consideration. Are we on the list?
- Presence across a fixed prompt set, by assistant
- Accuracy of how you're described, checked against your positioning
- Which competitors you're shortlisted alongside
- Share of self-reported sources naming AI, peers, or reviews
Tier two: Demand. Is consideration turning into interest?
- AI referral sessions, by assistant
- Inbound volume and cost per lead, by self-reported source
- Branded search trend, read as context rather than credit
Tier three: Revenue. Is interest turning into pipeline?
- Qualified pipeline by source, including AI-influenced deals logged by sales
- Sales cycle length for AI-influenced deals against other cohorts
- Win rate, same comparison
- Conversion rate of AI referral traffic, judged against your own baseline
Three rules make it usable:
Report the tiers together. Consideration metrics on their own look like vanity numbers. Revenue metrics on their own can't explain why pipeline moved. The value is in how they move together over time, reviewed on a standing cadence so you can see what's working, cut what isn't, and feed the learnings into the next cycle.
Reset expectations on precision. Tell leadership plainly: the click-based dashboard was precise but increasingly incomplete. It measured one event cleanly while missing a growing share of how buyers actually buy. The new one is less precise and more honest. That conversation is far easier before the quarterly review than after it.
Stop targeting click-through rate. Keep tracking it as a diagnostic. Stop setting goals on it. A team held to a click-through target on queries where AI answers appear is being measured on something it doesn't control, and it will make poor decisions to protect the number.
For a CRO, this ties marketing to pipeline predictability. For a CEO, it answers a more basic question: is the market hearing a consistent story about us in the places buyers now form opinions?
The takeaway
The click told you three things at once: that a buyer found you, that they cared, and where they came from. AI answers have split those signals apart, and no native report puts them back together.
So measure each one on its own terms. Consideration through presence and accuracy. Demand through referrals, inbound, and what buyers tell you. Revenue through pipeline, velocity, and win rate. Read every external benchmark against its own sample before it touches your plan.
Teams that make this shift now will have a year of comparable data by the time their competitors notice the dashboard stopped telling the truth. Measurement built for how buyers buy today compounds. Measurement inherited from the click era doesn't.
If you’re rethinking how your team measures marketing, let’s talk.
Dan
Managing Director, Slanted Frame. Writes about B2B growth systems, AI in marketing, and content marketing.