Discover how AI search and invisible buyer journeys are changing marketing attribution, brand visibility, and performance measurement.
If your attribution reports (the reports that assign credit for a sale to the marketing touches that came before it) have started to feel less reliable, your read on them is most likely correct. For about twenty years, the core promise of digital marketing was that you could see the whole path: who clicked, what they read, where they came from, and what they did next, which is why attribution became the discipline, the funnel became the model, and measurement became the pitch.
That visibility is shrinking, and a better analytics platform will not restore it, because more of the buying journey now happens in places that never report back to you, such as AI assistants, answers shown directly on the results page, private messages, and voice queries. Buyers have not stopped researching. They are doing more of it out of your view.
Most marketing operations still assume the journey can be watched. This post lays out a different approach for leadership: stop trying to track every step, and put your effort into shaping the information those unseen steps draw on.
The invisible journey
Stop tracking the journey. Shape what it draws on.
-
You can’t see
A buyer’s conversation with an AI assistant
You can shape
What that assistant knows about you
-
You can’t see
A buyer reading an AI answer
You can shape
Whether you are the source it cites
-
You can’t see
A colleague’s private recommendation
You can shape
The reputation that makes it land
Measurement becomes inference
-
Share of search
Your share of branded searches in your category, against competitors
-
Branded and direct
Where good invisible journeys show up first
-
AI answer sampling
Monthly checks of how AI describes you
-
“How did you hear about us?”
Customers’ own words catch what analytics can’t
of AI-cited sources for consumer goods are brand websites
of AI search users double-check against other sources
Sources: McKinsey, State of the Consumer 2026. Morning Consult for Yelp, April 2026.
How much of the buying journey is now out of view
The part of the journey you can see is now the smaller part.
In the first four months of 2026, 68% of U.S. Google searches ended without a click to any website, according to SparkToro's analysis of Similarweb data, which means about two in three searches now finish on Google's own page (these are often called zero-click searches). McKinsey reported in October 2025 that about half of Google searches already include an AI summary at the top of the results, and it expects that share to pass 75% by 2028, so the gap between searches and site visits is likely to keep widening.
Search is only one of the channels moving out of view. McKinsey's AI Discovery Survey of 1,927 U.S. consumers found that 44% of people who use AI-powered search call it their primary and preferred source of insight, ahead of traditional search at 31%, retailer or brand websites at 9%, and review sites at 6%, which puts the brand website near the bottom of the list buyers turn to first. About half of consumers now use AI-powered search on purpose to guide purchase decisions, including a majority of baby boomers, so the shift reaches well beyond younger buyers. McKinsey projects that $750 billion in U.S. revenue will flow through AI-powered search by 2028, a forecast large enough to belong in your budget planning.
The effect on measurement is direct. When a buyer asks an AI assistant which vendors to consider, you see none of that conversation, and when an AI answer resolves the question without a click, you see none of that either. When a colleague forwards a recommendation in a private message, it also stays out of your view. Add the blind spots that have always existed, such as word of mouth, dark social (sharing through private channels like text and messaging apps), and untracked referrals, and the part of the journey you can track becomes the minority.
Why the measurement-first playbook no longer works
Optimizing what you can see is no longer the same as optimizing what drives the decision.
The measurement-first approach was reasonable for its time. For most of the digital era, the visible journey was most of the journey, clicks mapped fairly well to behavior, and attribution, while imperfect, came close enough to reality to guide budget decisions.
That is no longer true, and the larger risk now is optimizing the fragment you can track while the deciding part of the journey happens somewhere else. Last-click attribution (a model that gives all the credit to the final touch before a sale) over-credits that last measurable step and under-credits everything that built the buyer's conviction before they typed your name. A team managing only the visible slice is, in essence, polishing the one room it can see in a house it can no longer walk through.
The natural instinct is to ask for better tracking. But AI search has removed much of the transparency that digital marketing was built on, and it is unlikely to come back, so the more useful question for leadership is what this loss of visibility requires your team to change.
How to influence the steps you cannot see
When you cannot watch the journey, you can still influence the information it draws on.
You cannot read a buyer's conversation with an AI assistant, but you can influence what that assistant knows about you by making your own information clear and by building enough credible outside coverage that the description a buyer receives is accurate and favorable. You cannot see a buyer read an AI answer, but you can work to become a source it cites. You cannot track the private message where someone recommends a vendor, but you can build the reputation that makes that recommendation go your way.
The approach is the same each time: instead of trying to observe the touchpoint, improve the material it draws on, and most of that material sits outside your own site. McKinsey's State of the Consumer 2026 report found that only 1% of the sources AI tools cite for consumer goods questions are brand-owned websites, which means nearly everything those tools say about a brand comes from third parties the brand does not control. Buyers also check what AI tells them, since a Morning Consult survey for Yelp found that 63% of AI search users double-check results against other trusted sources such as news sites and review platforms.
This raises the value of work the measurement era underfunded because it was hard to attribute: earned media, analyst coverage, customer reviews, community discussion, and clear, well-structured information about your company. These were once credibility signals for human readers, and they are now most likely also the inputs AI tools use when a buyer asks who to trust. When an AI tool is often the one describing your company to a buyer, the evidence that backs up what you say about yourself matters more than any ad you buy.
You cannot track the invisible journey, but you can shape what it draws on.
Measurement does not disappear. It becomes inference.
You cannot count what you cannot see, but you can read the signals it leaves behind.
Share of search. Your brand's share of searches in your category, compared with competitors, shows how present you are in journeys you cannot observe. A rising share suggests more of the unseen research is going your way.
Branded search and direct traffic. When someone searches your company name or types your web address directly, an unseen journey most likely led them there, which is why these numbers are often the first place a strong unseen journey shows up.
AI answer sampling. Asking AI tools the questions your buyers ask on a regular schedule, and recording how the tools describe you and your competitors, gives you a read on conversations you cannot see directly.
Self-reported attribution. Asking new customers how they found you, in their own words, catches what no analytics platform will, such as the AI recommendation, the podcast mention, or the colleague's message.
None of these signals is as clean as a click report, but each one describes what is happening in your market, which is more than a last-click model can claim. We cover specific metrics, including cost avoidance and citation share, in "The Day Google Stopped Asking What You Wanted" and "The CMO's Guide to the New Search Era."
The numbers behind the shift
AI search as the front door
About half of Google searches already show an AI summary. McKinsey projects $750 billion in US revenue will flow through AI search by 2028.
44% of AI search users already call it their primary source of insight. Most still double-check what it tells them, and brand websites supply little of what it cites.
| Finding | Figure | Source |
|---|---|---|
| Measured now | ||
| Google searches that show an AI summary | ~50% | McKinsey, October 2025 |
| US Google searches ending without a click, January to April 2026 | 68% | SparkToro, Similarweb data, June 2026 |
| AI search users who call it their primary source of insight, versus traditional search | 44% vs 31% | McKinsey AI Discovery Survey, 1,927 respondents, October 2025 |
| AI search users who double-check results against other trusted sources | 63% | Morning Consult for Yelp, April 2026 |
| Brand-owned websites as a share of AI citations, consumer goods | 1% | McKinsey, June 22, 2026 |
| Projected to 2028 · McKinsey | ||
| Google searches expected to show an AI summary by 2028 | 75%+ | McKinsey, October 2025 |
| US revenue projected to flow through AI-powered search by 2028 | $750 billion | McKinsey, October 2025 |
The 2028 figures are McKinsey forecasts, not measurements. SparkToro’s zero-click figure covers US searches and excludes searches in Google’s mobile app. Yelp commissioned the double-check survey.
Survey figures describe what people report; the revenue figure is a projection.
What the shift costs and what it returns
Acting on inference is uncomfortable, but waiting for certainty usually costs more.
Moving from tracking to shaping costs some comfort. Leadership has to accept signals that show a direction instead of an exact count, budgets move toward work that is harder to attribute, such as PR, reviews, and content quality, and the payoff takes patience because reputation builds slowly.
In return, you gain influence over the part of the journey that most likely decides outcomes, along with a measurement approach that covers the whole journey, including the parts you cannot track. It is also worth doing now because many competitors are still waiting for a level of precision that AI search has removed, and that delay gives your team time to build a lead.
Four failure modes to watch for
Each one feels responsible, and each one keeps a team optimizing the wrong thing.
The first is demanding pixel-level proof (tracking tied to a specific code on a web page) before acting, which guarantees a team only ever optimizes the visible minority. The second is cutting brand and PR budgets because they do not show up in attribution, which removes the very inputs AI tools draw on. The third is treating AI answer checks as a one-time audit instead of a recurring measure, even though those answers may change from month to month. The fourth is collecting self-reported attribution and then ignoring it because it does not match the dashboard, and when the two disagree, the customer's own account is usually closer to what happened.
Test small, then make it a routine
Prove the new signals are useful before you restructure your reporting.
Start with three to five of the questions your buyers ask most, which is a practical starting size, not an official benchmark. For one quarter, sample how AI tools answer those questions each month, track branded search and direct traffic, and add a one-question "How did you hear about us?" field to your onboarding. Then compare what these signals show with your attribution reports, and if they reveal channels your dashboard misses, expand them, and if they do not, you have spent very little to find out.
Once the test proves useful, write the routine on one page and hand it to a colleague. If they can run it next quarter without you, the practice will hold up through staff changes and busy seasons.
What this likely means over the next 6 to 18 months
Expect the share of the journey you cannot see to keep growing.
If current trends continue, AI summaries will most likely appear on more searches, more buyers will start their research in AI assistants, and more of the journey will happen in places that do not report back. The companies in the strongest position will probably be those that have already moved budget toward the inputs AI tools rely on and adopted measures that read demand they cannot see directly, while those still waiting for perfect attribution are likely to keep optimizing a shrinking share of the journey.
To see how AI tools describe your brand today, Camino5's free AEO Invisibility Score shows you in about 30 seconds and sends the results by email.
How to start today
Ask four AI tools what they say about your brand, your category, and your top competitors, and record the answers. Repeat monthly.
List the publications, review sites, and communities AI tools cite most for your category, and make earning accurate coverage there part of next quarter's plan.
Add a one-question "How did you first hear about us?" field to your onboarding or checkout, with a free-text answer.
Put branded search, direct traffic, and share of search on your monthly dashboard next to your attribution reports.
Agree with leadership on which directional signals you will act on, so the team does not wait for proof that will not come.
Most of the buying journey now happens where you cannot see it; the brands that win will be the ones that stop trying to watch it and start shaping what it is built from.
Sources
McKinsey & Company, "New front door to the internet: Winning in the age of AI search," October 2025
Search Engine Land, "Google zero-click searches hit 68% in early 2026: Study," 2026
McKinsey & Company, "State of the Consumer 2026," June 22, 2026
Yelp, "Americans Use AI But Don't Trust It" (survey with Morning Consult), April 2026