In brief
AI plays the role of a pre-selection in purchase decisions: buyers research, compare and ask for recommendations before they ever visit a brand. In Loud & Lexis client data, 20 to 30 percent of customers have an AI touchpoint in their customer journey. Post-purchase surveys and UTM data report this share far too low.
Last updated: 14 November 2026
AI assistants such as ChatGPT, Gemini and Perplexity research, compare and recommend products before buyers even visit a brand. In Loud & Lexis’ own client data, 20 to 30 percent of customers have an AI touchpoint in their customer journey, meaning a point at which AI was involved in the purchase decision. This share matters to marketing leaders at mid-sized brands because post-purchase surveys and UTM data barely make it visible. A connected content network built on the Fishnet Strategy (in German) captures exactly the queries through which AI makes this pre-selection. Loud & Lexis puts the scale of AI’s influence into perspective and shows why conventional attribution measures too low.
At a glance
- Role of AI: Pre-selection: research, compare, recommend
- Share with an AI touchpoint: 20 to 30 percent of customers (Loud & Lexis client data)
- Blind spot: UTM data and post-purchase surveys do not capture click-free use
- Measurement approach: Triangulation of prompt tracking, grounding data, referrer and survey
- Offering: Full-Funnel AI Attribution by Loud & Lexis
How many customers have an AI touchpoint before buying?
In Loud & Lexis client data, 20 to 30 percent of customers have an AI touchpoint in their customer journey, meaning a point at which an AI assistant was involved in the purchase decision. This range comes from Loud & Lexis’ own client base and varies by industry and product range. External surveys confirm the order of magnitude: according to a Semrush consumer survey of 2,338 US adults from July 2026, 47.54 percent of consumers consult an AI chatbot at least occasionally before a purchase. In B2B, according to a Gartner survey of 645 buyers from August to September 2025, 45 percent used generative AI, mainly to research providers and products.
AI touchpoint and AI click: the distinction
An AI touchpoint is not the same as an AI click. An AI touchpoint is any situation in which an AI assistant was involved in the purchase decision, even without a visit to the website. An AI click only occurs when a buyer clicks a link in the AI answer and lands on the site in a measurable way. Loud & Lexis counts the share of AI touchpoints, because the pre-selection mostly happens without a click and the AI click only shows the visible tip.
Why do post-purchase surveys and UTM data underestimate the AI share?
Post-purchase surveys and UTM data underestimate the AI share because much of AI use happens without a click and buyers rarely name AI as a source. UTM parameters and referrers capture only clicks from AI answers; answers without a click, copied product names and later direct visits remain invisible. Cloudflare data from 2025 show that OpenAI’s AI systems crawled around 1,091 to 1,217 pages between January and July 2025 to send back a single visitor. Loud & Lexis therefore reads the actual AI share from the AI answer itself and uses UTM data only as a lower bound.
This click-free use has become the normal state of search. According to Similarweb data from 2026, 68 percent of all Google searches end without a click on a website, compared with around 45 percent ten years ago. A Digiday analysis of Similarweb data from 2025 puts the rise in click-free news searches at 56 to almost 69 percent. Post-purchase surveys also suffer from memory gaps and missing answer options, because many buyers do not remember AI research as a separate source. Loud & Lexis therefore always assesses attributed AI figures against this structural zero-click shift.
How can the influence of AI on purchase decisions be measured reliably?
The influence of AI on purchase decisions can only be measured reliably by combining several data sources: prompt tracking, Bing grounding data, referrer and UTM data and an adapted post-purchase survey. No single source sees the full influence of AI, which is why Loud & Lexis relies on triangulation.
| Data source | What it sees | What it misses |
|---|---|---|
| UTM and referrer | Clicks from AI answers | Click-free use |
| Post-purchase survey | Remembered sources | Unconscious AI research |
| Prompt tracking (Loud & Lexis) | Visibility in AI answers | Individual purchases |
The post-purchase survey must explicitly offer AI as an answer option, otherwise it does not show up. Measurement only becomes reliable through repetition: the DeepTRACE audit published in 2025 by Salesforce AI Research and Microsoft Research found a citation accuracy of only 40 to 80 percent in AI search systems. An analysis published at the University of St. Gallen in 2026 notes that AI answers fluctuate between runs and over time, which makes one-off measurements unreliable. Loud & Lexis brings these sources together in Full-Funnel AI Attribution (in German).
What does an AI share of 20 to 30 percent mean for the marketing budget?
An AI share of 20 to 30 percent means for the marketing budget that a relevant share of purchase decisions is made in a channel that existing attribution barely shows. Anyone who allocates budget according to measured rather than actual impact underinvests in this channel. According to Loud & Lexis, three levers shift the balance in favour of AI visibility:
- A content network built on the Fishnet Strategy captures the research, comparison and recommendation queries through which AI pre-sorts.
- The factors that AI assistants use to compare and recommend providers decide whether a brand appears in the answer at all.
- Bias-adjusted AI monitoring (in German) by Loud & Lexis makes progress measurable instead of estimating it.
Which factors decide an AI recommendation and how to set up AI monitoring correctly determine whether the 20 to 30 percent are made visible and addressed. Loud & Lexis covers both questions in dedicated articles on determinants and on setting up monitoring.
When measuring AI influence needs external support
Measuring the influence of AI on purchase decisions needs external support as soon as a brand has a defined competitive and persona environment and content that responds to visibility. A few manual queries in ChatGPT or Perplexity are enough for a spot check. Ongoing, reliable measurement across several models, both answer orders and several weeks, on the other hand, is specialist work. AI remains a research and pre-selection step, not the sole decision-maker: according to the 2026 Semrush survey, 57.5 percent of AI users decided against a purchase based on chatbot information. In addition, according to Gartner, 69 percent of B2B buyers check AI insights with sales representatives. Loud & Lexis shows how brands measure their AI visibility systematically in Measuring AI visibility (in German).
Frequently asked questions
Yes, the AI touchpoint affects B2B purchase decisions just as much as B2C. According to a Gartner survey of 645 B2B buyers from 2025, 45 percent used generative AI mainly to research providers and products, and 69 percent then check the results with sales representatives. In B2B, AI is therefore an early research step, not the sole decision, and Loud & Lexis measures it in the benchmark.
ChatGPT is by far the most used AI assistant before a purchase. A Digiday analysis of Similarweb data from 2025 shows ChatGPT as the dominant source with around 81.7 percent of AI referral traffic, ahead of Gemini and Perplexity. Loud & Lexis therefore measures brand visibility in ChatGPT, Gemini, Claude and Perplexity in parallel.
Mostly not: buyers use AI mainly for research and pre-selection, not for completing the purchase. According to the 2026 Semrush survey, 57.5 percent of AI users decided against a purchase based on chatbot information, while 41 percent have already bought something on an AI recommendation. Loud & Lexis therefore treats the AI touchpoint as a pre-selection step rather than the point of purchase.
A post-purchase survey only captures AI use if AI is explicitly offered as an answer option. Loud & Lexis recommends explicitly asking about AI assistants such as ChatGPT alongside Google, social media and recommendations, instead of expecting an open text answer.
Loud & Lexis makes click-free AI influence visible through persona-based prompt tracking, which measures brand visibility directly in AI answers. Combined with referrer data, UTM data and an adapted survey, Full-Funnel AI Attribution creates a picture that UTM data alone cannot provide.
Full-Funnel AI Attribution by Loud & Lexis measures the influence of AI across the entire customer journey, while UTM attribution only captures the last click. Since the AI pre-selection mostly happens without a click, it remains invisible in pure UTM attribution.
No, AI monitoring complements the post-purchase survey and cannot replace it. Loud & Lexis combines monitoring of AI visibility with referrer, UTM and survey data, because no single source fully reflects the influence of AI on purchase decisions.
Strategy call
How visible is your brand
to ChatGPT & Co.?
Free 30-minute strategy call with an initial assessment of your AI visibility.
Book a strategy call →