AI Discovery Is Hiding the Middle of the Customer Journey. 86% of Advertisers Are Rebuilding Measurement.

On September 10, the Interactive Advertising Bureau reported that 86% of surveyed ad buyers are changing—or expect to change—media measurement because of conversational AI and AI agents.

The commercial consequence is immediate: a customer can discover, compare and shortlist brands inside an AI answer before producing any click a marketing system can observe.

The decision businesses must reconsider is whether click-path attribution still deserves to be treated as the complete account of how demand was created.

The IAB’s September update surveyed more than 200 brand and agency advertising decision-makers. It is an industry outlook, not an audited census of every advertiser, but the direction is unusually clear. AI is not only changing campaign production. It is removing visible steps from the customer journey faster than measurement practices can replace them.

The missing middle is becoming normal

Traditional digital measurement assumes a trail. A customer sees an ad, visits a page, returns through search, opens an email or converts through a platform that claims the final interaction. The trail has always been incomplete, but marketers could usually observe enough of it to construct a useful sequence.

Conversational discovery compresses that sequence. A person can ask an AI system for category explanations, alternatives, price logic, objections and a final shortlist without visiting the brands being discussed. When that person eventually searches for one company directly or arrives at a retailer ready to buy, the visible journey may begin near the end.

The IAB found that 45% of buyers considered the difficulty of comparing AI-driven and traditional customer journeys their leading measurement challenge. That is not merely a reporting inconvenience. It means channels that create consideration can disappear from the record while the final destination receives too much credit.

The last click has become a weaker witness

Attribution models do not become useless because AI enters the journey. They become less qualified to answer questions they were never designed to solve. A click can still show how someone arrived. It cannot reliably reveal where the person formed a preference, rejected an alternative or learned the language that made the purchase feel safe.

This changes the leverage inside a marketing report. Paid search, direct traffic and retail conversion pages may appear stronger because they capture customers after AI-assisted consideration. Meanwhile, useful articles, creator recommendations, expert mentions and distinctive brand signals may look weaker because their influence was summarized by a system rather than clicked by a person.

The danger is not simply undercounting content. It is reallocating money toward the most visible endpoint and gradually starving the sources that made the endpoint productive.

Visibility inside AI is becoming a media concern

The IAB found that adapting to AI-driven search was the top media-investment challenge for 44% of buyers. Optimizing content for AI-generated answers was the leading area of increased focus, selected by 76%. Those figures suggest that AI visibility is moving beyond an SEO experiment and into the planning conversation traditionally reserved for paid media, creators and publishers.

The leverage shifts toward businesses with strong brand meaning, credible source material and enough first-party knowledge to recognize changes in demand. Creators and publishers with trusted audiences also gain strategic importance because AI systems can absorb and restate the opinions customers use to decide.

Brands that depend on generic category language lose leverage. If an AI answer can describe several competitors interchangeably, producing more content may only make the category easier to summarize without making one company easier to choose.

Measurement now needs several witnesses

According to the IAB, 48% of buyers are measuring brand visibility and citations inside AI tools. Others are watching branded search, direct traffic, AI referrals and incrementality. None of those signals is a perfect replacement for the missing click path. Together, they can provide a more defensible explanation than any single dashboard.

The practical shift is from attribution certainty to evidence triangulation. A business can compare whether AI visibility changes, branded demand rises, direct visits behave differently and controlled campaigns create incremental outcomes. The goal is not to assign every sale to one source. It is to learn which combination of signals consistently precedes stronger business results.

That approach also demands memory. Without a record of what changed in the strategy, content, campaign and customer response, a new signal is merely another number.

Keep the business objective visible

OrionPilot’s relevance is clearest at this point of uncertainty. Keeping weekly direction, content work, campaign support and performance learning connected around the business objective can help an operator compare incomplete signals without letting the easiest metric dictate the next move. OrionPilot does not make the hidden journey fully observable; it helps preserve the strategic context needed to learn from what can be observed.

Businesses should therefore reconsider measurement before adding more AI-generated activity. Define which changes would count as stronger discovery, what independent outcome should confirm them and which signals deserve attention even when they cannot claim a conversion.

The IAB’s finding is not that measurement has ended. It is that the customer journey has moved into environments the old trail cannot fully enter. The next advantage belongs to businesses that can make decisions from several credible witnesses without pretending any one of them saw everything.