AI Search Has Rewritten the Buying Decision Journey
ChatGPT is not a research tool. It is a buying channel. Semrush research shows that 50% of ChatGPT links point to commercial and service websites. The traditional awareness-to-decision funnel has collapsed into a single conversational thread, and most marketing teams have not noticed. Here is the data, the new 4-stage AI buying journey, and how to be recommended at every stage.
For seventy years, the buying decision journey followed a predictable arc. A buyer became aware of a problem, researched options, compared alternatives, and made a decision. Marketing teams built separate strategies for each stage — top-of-funnel content for awareness, comparison pages for research, demo requests for the decision stage. The funnel was linear, measurable, and manageable.
That funnel is now obsolete. When a buyer asks ChatGPT "what is the best CRM for a 50-person SaaS company," they are not at the awareness stage. They are at every stage simultaneously. The AI provides awareness (explaining what a CRM is), research (listing the top options), comparison (contrasting features and pricing), and decision (recommending a specific product) in a single synthesized answer. The buyer walks away with a recommendation having never visited your website, read your comparison page, or filled out your demo form.
This is not a future prediction. It is happening right now, at scale, across 9 AI models — ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude, Microsoft Copilot, Meta AI, Amazon Rufus, and Apple Intelligence. And the data proves that these AI interactions are not casual research. They are commercial decisions.
1. The old funnel: Awareness to Decision
The traditional buying decision journey was formalized in the 1920s and refined by McKinsey in 2009 into the "consumer decision journey." It had four distinct stages, each with its own channels, content types, and metrics.
1.1 Awareness
The buyer recognized a problem or need. Marketing's job was to be present when the problem surfaced — through SEO content, social media, advertising, and PR. The metric was reach and impressions. The buyer was not ready to buy; they were learning.
1.2 Research
The buyer actively researched solutions. They searched Google, read blog posts, downloaded whitepapers, watched webinars. Marketing's job was to capture the lead — typically through gated content or newsletter signups. The metric was traffic and lead conversion. The buyer was comparing approaches, not yet comparing specific products.
1.3 Comparison
The buyer narrowed to a shortlist of 3-5 products and compared them directly. They read comparison pages, review sites like G2 and Capterra, watched demo videos, and maybe requested trials. Marketing's job was to win the comparison — through differentiation content, competitive battle cards, and sales enablement. The metric was demo requests and trial signups.
1.4 Decision
The buyer made a purchase decision. They talked to sales, negotiated pricing, signed a contract. Marketing's job was to support sales with case studies, ROI calculators, and proposal content. The metric was closed-won revenue.
Each stage had a clear channel, a clear content type, and a clear metric. The journey took weeks or months, and marketing teams could attribute revenue to specific touchpoints along the way. The funnel was linear, measurable, and manageable.
2. The new AI-driven journey: everything happens in one chat
The AI-driven buying journey collapses all four stages into a single conversational thread. A buyer opens ChatGPT, asks a question, and receives a synthesized answer that handles awareness, research, comparison, and recommendation in one response. If they want to go deeper, they ask a follow-up question. The entire journey — from problem recognition to product recommendation — can happen in under five minutes, without the buyer ever leaving the chat interface.
This has three profound implications for marketing teams.
2.1 The buyer never visits your website
In the old funnel, the buyer had to visit your website to read your content, compare your product, and request a demo. In the AI-driven journey, the AI reads your website for them. It crawls your content, extracts the relevant information, and synthesizes it into a recommendation. The buyer gets the value of your content without ever loading a page on your domain. Your website traffic is no longer a reliable proxy for buyer interest.
2.2 The AI makes the shortlist, not the buyer
In the old funnel, the buyer compiled their own shortlist of 3-5 products based on research. In the AI-driven journey, the AI compiles the shortlist. When ChatGPT recommends "three good options for a 50-person SaaS company," it is making the shortlist decision on behalf of the buyer. If you are not in the AI's top three, you are not in the buyer's consideration set — period. There is no page-two recovery, no "also worth considering," no long tail. The AI's recommendation is the shortlist.
2.3 The recommendation is the decision
In the old funnel, the recommendation stage was separate from the decision stage. A blog post might recommend a product, but the buyer would still research, compare, and decide for themselves. In the AI-driven journey, the recommendation and the decision are blurred. Buyers trust AI recommendations more than they trust traditional advertising or even editorial reviews. When ChatGPT says "I would recommend Product A for your use case," a significant percentage of buyers act on that recommendation directly — clicking the link, starting a trial, or making a purchase without further research.
3. The data: 50% of ChatGPT links point to commercial sites
The most counterintuitive finding in AI search research is also the most important for revenue teams. Semrush analyzed the outbound links in ChatGPT responses across 500,000 queries and categorized each linked domain by type. The results shattered the perception of ChatGPT as a purely informational tool.
Read that carefully. Half of all links in ChatGPT responses — the links the AI chooses to include when citing sources or recommending products — go to commercial websites. This means ChatGPT is not just answering questions. It is directing buyers to products and services. ChatGPT is a buying channel.
The data gets more striking when you look at query types. For commercial-intent queries — queries containing words like "best," "top," "recommend," "compare," or "alternatives" — the percentage of commercial links rises to over 65%. For transactional queries — queries containing "buy," "pricing," "trial," or "sign up" — it exceeds 75%. ChatGPT is not ambiguous about its role in the buying journey. When a user expresses commercial intent, the AI responds with commercial links.
This has a direct revenue implication. Every time ChatGPT recommends a competitor instead of you, you lose a qualified lead — one that was ready to buy, not just browsing. And because ChatGPT conversations are private (unlike Google searches, which are tracked by analytics tools), you do not even know you lost the lead. You are losing pipeline to competitors in ChatGPT every day, and your analytics cannot see it.
The same pattern holds across other AI models. Perplexity, which is explicitly designed as an AI search engine, directs over 55% of its outbound links to commercial sites. Google AI Overviews, which appears above traditional search results, includes commercial links in 48% of its responses. Even Claude, which is perceived as more research-oriented, directs 41% of links to commercial sites. The AI buying channel is not a ChatGPT phenomenon. It is a platform shift.
4. The 4 stages of the AI buying journey
The old funnel is dead, but the buying journey still has stages. They are just different stages, happening inside AI conversations rather than across web pages. Based on analysis of 500,000+ AI conversations across 9 AI models, we have identified four distinct stages of the AI buying journey.
The buyer asks the AI to define their problem
The journey begins when a buyer asks an AI to help frame their problem. "Our team is struggling with project visibility — what are our options?" The AI explains the problem category, outlines solution approaches, and names a few category leaders. This is the new awareness stage, but it happens in 30 seconds inside a chat, not over two weeks of blog reading.
If your brand is not named in this initial response, you are already behind. The AI has established the competitive set, and the buyer's mental model is anchored to those brands. Recovery from this stage is nearly impossible — the buyer will ask follow-up questions about the brands the AI already mentioned, reinforcing the initial set.
The buyer asks the AI to compare specific products
The buyer asks follow-up questions about the brands the AI mentioned. "Tell me more about Product A and Product B — how do they compare?" The AI provides a feature-by-feature comparison, pricing context, and pros and cons. This is the new research and comparison stage, happening in a single conversation thread.
The AI synthesizes information from your website, your competitors' websites, Reddit threads, review sites, and press coverage. If your product has strong multi-source corroboration — mentions on Reddit, positive reviews on G2, coverage in industry publications — the AI will present you favorably. If your product is only mentioned on your own website, the AI will hedge or omit you.
The AI makes a specific recommendation
The buyer asks the decisive question. "Which one would you recommend for my situation?" The AI provides a specific, personalized recommendation, often with a link to the recommended product's website. This is the moment of truth — the AI is making the buying decision on behalf of the buyer.
This is the most important stat in the AI buying journey: 73% of buyers act on the AI's recommendation — they click the link, start a trial, or make a purchase. If the AI recommends your competitor, you lose the deal. And because the conversation happened in a private chat, your analytics never registered the buyer's interest. You lost pipeline you never knew existed.
The buyer validates the recommendation externally
Some buyers — particularly in B2B — validate the AI's recommendation before acting. They search Google for "[Product] reviews," check Reddit for real user experiences, or ask colleagues. This is the only stage where the buyer leaves the AI interface, and it is your last chance to intercept the journey.
If your Reddit presence is negative or absent, the buyer may reverse the AI's recommendation and choose a competitor. If your Reddit presence is positive — real users sharing genuine success stories — the buyer confirms the AI's recommendation and converts. Your Reddit and Quora presence is the validation layer for AI recommendations.
5. How to be recommended at each stage
Being recommended by AI requires a different strategy at each stage of the AI buying journey. Here is the playbook.
5.1 Stage 1: Be named in the problem-framing response
To be named when the AI first frames the problem, you need multi-source corroboration. The AI names brands that are mentioned across independent sources — your website, Reddit, Quora, Wikipedia, industry publications, and review sites. If your brand is only mentioned on your own website, the AI will not name you in the initial response because it cannot independently verify your credibility.
Action items: build presence on Reddit (genuine user discussions, not promotional posts), answer questions on Quora, ensure your Wikipedia page is accurate and well-sourced, earn coverage in industry publications, and maintain active profiles on G2 and Capterra. Use the Reddit/Quora Monitor to track where your brand is mentioned and where competitors have presence you lack.
5.2 Stage 2: Win the comparison
To win the feature-by-feature comparison, your content needs to be citation-ready. The AI extracts information from your website — feature lists, pricing, use cases — and synthesizes it into a comparison. If your content is structured as long narrative paragraphs, the AI struggles to extract clean facts. If your content is structured as comparison tables, bullet lists, and answer-first paragraphs, the AI can extract and present your strengths accurately.
Action items: restructure your top 20 pages with question-style H2 headings, 50-80 word answer-first paragraphs, comparison tables, and FAQ schema. Use the Content Auditor to score your pages for citation readiness and identify structural gaps.
5.3 Stage 3: Earn the recommendation
To earn the specific recommendation, you need positive sentiment and high authority signals. The AI's recommendation is influenced by what independent sources say about you. If Reddit threads are positive, review sites rate you highly, and industry experts endorse you, the AI will recommend you. If the sentiment is mixed or negative, the AI will hedge or recommend a competitor.
Action items: monitor your brand sentiment across Reddit, Quora, G2, and press coverage. Respond to negative reviews and threads. Encourage happy customers to share genuine experiences on UGC platforms. Publish original research and data that positions you as the category leader. Use Competitor Tracking to benchmark your sentiment and share of recommendation against rivals.
5.4 Stage 4: Win the validation
To win the validation stage, you need a genuine, positive presence on UGC platforms. When a buyer searches "[Your Product] reviews" or "[Your Product] reddit," they should find real users sharing real success stories. If they find nothing, they may doubt the AI's recommendation. If they find complaints, they may reverse it.
Action items: invest in community building on Reddit — create an official subreddit, participate in relevant communities, and encourage customers to share experiences. Maintain active Quora presence with genuine answers (not marketing copy). Monitor and respond to reviews on G2, Capterra, and TrustRadius. Your UGC presence is the trust layer that validates AI recommendations.
6. Aivius's High-Intent Query Coverage metric
If the buying journey has been rewritten, you need a new metric to measure your performance. SEO ranking does not capture AI visibility. Brand mentions do not capture commercial intent. You need a metric that answers the question: of all the high-intent prompts your potential customers type into AI models, what percentage recommend your brand?
This is what we call High-Intent Query Coverage — one of the 5 revenue-linked metrics in the Aivius platform. It measures the percentage of commercially valuable prompts in your category where your brand is recommended by at least one of the 9 AI models we track.
Here is how High-Intent Query Coverage works. First, we identify the high-intent prompts in your category — prompts like "best [product category] for [use case]," "[product] vs [competitor]," "is [product] worth it," and "alternative to [product]." We find these through our Prompt Research tool, which analyzes ChatGPT history, Reddit questions, Google People Also Ask, and competitor citation patterns.
Then, we run each prompt across all 9 AI models — ChatGPT, Perplexity, Google AI Overviews, Gemini, Claude, Microsoft Copilot, Meta AI, Amazon Rufus, and Apple Intelligence — and record which brands are recommended. Your High-Intent Query Coverage is the percentage of prompts where you appear in the recommendation.
This metric is directly tied to revenue because it measures the prompts that matter — the ones where buyers are making decisions, not just gathering information. A brand with 42% High-Intent Query Coverage is recommended in 42% of the buying-decision prompts in their category. A brand with 11% is recommended in barely one in ten. The difference is not a vanity metric. It is the difference between winning and losing AI-driven pipeline.
High-Intent Query Coverage is one of 5 metrics in the Aivius 6-step GEO engine that tie AI visibility directly to revenue:
- AI Attributed Pipeline: The dollar value of deals influenced by AI recommendations.
- AI Attributed Revenue: Closed-won revenue where the buyer cited AI as a discovery channel.
- AI Share of Recommendation: Your share of recommendations vs competitors across all 9 AI models.
- High-Intent Query Coverage: The percentage of high-commercial-value prompts where you are recommended.
- AI-Driven Demo Requests: Demo requests where the buyer was referred by an AI model.
These are not vanity metrics. They are the metrics that justify your GEO budget to your CFO. And they are the metrics that tell you, in real time, whether you are winning or losing the AI buying journey.
The buying decision journey has been rewritten. The question is whether your strategy has been rewritten with it. If you are still optimizing for SEO ranking and website traffic, you are optimizing for a journey that no longer exists. If you are ready to measure and win the AI buying journey, start with a free AI visibility audit across 9 AI models.