As consumers increasingly use AI systems to research companies and products, businesses are confronting a new attribution problem: an AI recommendation can influence a purchase without generating the click that traditionally proved where the customer came from.
WHAT’S HAPPENING
For years, digital marketing followed a relatively visible path.
A customer searched Google, clicked a result or advertisement, visited a website and potentially made a purchase.
AI search is changing that journey.
A buyer can now ask ChatGPT, Gemini or another AI system for recommendations, discover a company inside the generated answer, remember the name and visit that company days later through Google, a direct website visit or even a paid advertisement.
The eventual conversion may be credited to that final channel even though AI originally introduced the brand.
HubSpot is now encouraging marketers to measure AI-search performance across three layers: visibility, engagement and revenue rather than relying solely on direct referral traffic. Its August 2026 guidance argues that perfect attribution is difficult because customers influenced by AI may later arrive through completely different channels.
The shift is already measurable.
Adobe reported that traffic from generative AI tools to U.S. retail websites increased 693.4% year over year during the 2025 holiday shopping season. Growth remained strong in the first quarter of 2026, when AI-driven retail traffic increased 393% year over year.
WHY IT MATTERS
This could create a marketing channel that companies know is influencing customers but cannot completely see.
Traditional digital advertising became valuable partly because clicks made attribution easier.
AI can remove that click.
A consumer may ask an AI assistant:
“What are the best companies for this?”
The AI may mention three businesses.
The customer researches one of them later and eventually purchases.
The company gets the sale.
But its analytics may never show that the first meaningful interaction happened inside an AI conversation.
That makes being included in the answer itself increasingly important.
HubSpot’s research found that 42% of CRM software buyers surveyed used AI search while evaluating vendors, and buyers who used AI search were 36% more likely to purchase than those who did not. That is a correlation rather than proof that AI caused the purchase, but it demonstrates how deeply AI can enter the evaluation process before a salesperson ever sees the customer.
WHO BENEFITS
Brands frequently recommended or cited by AI systems could gain awareness before customers ever reach a traditional search engine.
Companies with strong reputations and authoritative online information may benefit as AI systems evaluate which businesses and sources are credible enough to include in generated answers.
AI-search measurement platforms could benefit as companies look for ways to track mentions, citations, share of voice and AI-assisted conversions.
Consumers may benefit from faster research and the ability to compare businesses without navigating dozens of websites.
WHO LOSES
Brands that rank well in traditional search but rarely appear in AI-generated answers could lose visibility as more research moves inside AI systems.
Marketers dependent on last-click attribution may misunderstand which channels actually influenced a sale.
Companies with weak or inconsistent information across the web could become harder for AI systems to identify, describe or recommend accurately.
Traditional publishers and websites could also face pressure if consumers obtain enough information from AI-generated responses that fewer visits reach the underlying sites.
However, the transition should not be overstated.
A large analysis of more than 40,000 major U.S. websites found organic search traffic declined only 2.5% year over year, suggesting traditional search remains enormous even as AI grows rapidly from a much smaller base.
WHAT HAPPENS NEXT
Expect businesses to start tracking a new category of marketing indicators.
Not simply:
Where do we rank on Google?
But:
Does ChatGPT know us?
Does Gemini mention us?
Are we cited in AI answers?
How often do we appear compared with competitors?
Do branded searches increase after our AI visibility improves?
And ultimately:
Does appearing inside AI answers produce revenue?
Companies may increasingly measure what is being called Share of AI Voice — how frequently a brand appears across relevant AI-generated answers — alongside traditional SEO rankings, website traffic and advertising performance.
The larger transition is already beginning.
The internet is not abandoning search.
But consumers are gaining another layer between themselves and the websites they once visited directly.
And if AI increasingly decides which companies enter the conversation, the valuable position online may no longer simply be the first link.
It may be being part of the answer at all.
