Afternoon BriefAI Search & Discovery

Your Buyers Are Already Researching You in AI. The Proof Is in Your Bot Traffic.

Forrester says the B2B buyer journey has moved into AI answer engines — and the evidence is hiding in web traffic most marketers throw away. Here's the three-step audit to find it and act on it.

Christian Lehman|Mar 23, 2026

Your Buyers Are Already Researching You in AI. The Proof Is in Your Bot Traffic.

There's a metric most marketing teams are throwing away that contains the most useful buyer signal they're not reading.

Web security logs classify automated traffic into categories. Most teams discard it wholesale — bots, crawlers, noise. But a new Forrester analysis published in early March 2026 points to something operators need to hear: buried in that "bot traffic" bucket are real buyers using AI answer engines to research your company, compare you against competitors, and pre-qualify you before they ever fill out a form.

Those buyers didn't generate a click. They left no session data in your analytics. They didn't trigger a retargeting pixel. They were there, did their research, and moved on — invisible to your funnel but fully active in theirs.

If your pipeline has started to feel thinner than your traffic numbers suggest, this is part of why.

What Forrester's data actually shows

According to Forrester's Buyers' Journey Survey 2025, 94% of B2B buyers now use AI during their buying process. That number is near-universal at this point. But the more specific finding is what changes the operational picture: twice as many buyers named generative AI or conversational search as a more meaningful source of information than any other source, including vendor websites, product experts, and sales conversations.

Buyers are using AI tools — not as a supplemental search tool, but as their primary research interface. They're asking ChatGPT and Perplexity "who are the best [category] platforms for a [size/use case] company" before they ever open a browser tab to visit anyone's marketing site.

The traffic impact is real. Forrester documents B2B website traffic declines of 10–40% over the past year, tied directly to buyers shifting research into AI environments where you don't have visibility. Your impressions may be flat or rising. Your clicks are falling. This is the mechanism.

The critical insight Forrester added in its March 2026 analysis: when buyers use answer engines to research vendors, those interactions don't disappear — they show up in your server logs as bot traffic. ChatGPT, Perplexity, and Claude send recognizable crawler signatures. Security teams see them. Marketing teams don't, because the standard playbook has always been to discard automated traffic as noise.

Those automated visits aren't noise. They're buyer-assist agents — systems actively responding to a real human research question about your company.

The buyer prompted the AI, the AI dispatched a crawler to your site, the crawler retrieved information, and that information shaped whether you appeared in the answer and how you were described. All of that happened without generating a single session in GA4.

The three-step audit to find and use this data

Operators can act on this now, without waiting for the analytics industry to build AI-aware attribution infrastructure. The process is manual, but it takes under two hours the first time.

Step 1: Pull your bot traffic, classified by actor type

If you have access to your web server logs or your CDN's traffic analysis (Cloudflare, Fastly, etc.), filter for requests by user-agent strings associated with AI crawler systems. The key signatures to look for: GPTBot (OpenAI), PerplexityBot, ClaudeBot (Anthropic), Googlebot (for AI Overviews), and Bingbot (for Copilot). Most security tools or log analysis platforms can surface these directly.

Look at: which pages these crawlers visit, how often they visit, and what changed in crawler frequency over the last 90 days. The pages AI crawlers visit most frequently are the pages AI engines are using to form opinions about your brand. If your home page, product pages, and case studies are seeing regular AI crawler traffic, those are the pages shaping what ChatGPT or Perplexity says about you.

If your highest-traffic blog posts are not in that list, that's your first gap: AI engines may be building their model of you from a different set of pages than the ones you're investing in.

Step 2: Run the buyer prompt audit

Once you know which pages AI engines are crawling, run the corresponding buyer research prompts against ChatGPT, Perplexity, and Claude (with web search enabled). The prompts should mirror what your target buyer would actually ask — not branded queries, but category queries:

"What are the best [your category] platforms for a [company size] [industry] team?"
"Compare [you] vs [competitor 1] vs [competitor 2]"
"Who are the most credible [your category] vendors for [specific use case]?"

For each query: note whether you appear, where you appear, how you're described, and what sources are cited when you are mentioned. This is the output your bot traffic was feeding. Now you can see it directly.

If you're appearing inconsistently — cited in Perplexity but absent in ChatGPT, or described accurately in one and inaccurately in another — that tells you something about where your earned media coverage is concentrated and where the gaps are. ChatGPT draws primarily from training data accumulated over time; Perplexity uses real-time retrieval. A strong position in Perplexity but weak position in ChatGPT usually means you've built recent coverage but lack the sustained third-party presence ChatGPT's training cycle requires.

Step 3: Diagnose the citation gap and close it

The buyers researching you in AI aren't finding your best pages organically. They're asking a question and getting whatever AI has synthesized from the sources it already trusts. That synthesis is built from a specific input: earned media placements in publications AI engines have been trained to treat as authoritative.

According to the Muck Rack Generative Pulse analysis of over one million AI-cited links, more than 85% of non-paid AI citations come from earned media — third-party coverage in recognized outlets, not content on your own domain. The Fullintel-UConn academic study presented at the International Public Relations Research Conference in March 2026 found that 89%+ of all links cited by AI engines were earned media. Your blog post that does 5,000 visits a month from SEO is not competing with your TechCrunch mention in terms of AI citation weight.

Once you know which buyer queries your brand should be answering and which publications appear in AI citations for those queries, you have a prioritized target list. Not a generic media list — a citation map. The outlets showing up in AI answers for your category queries are the ones worth placing in. Getting into them systematically is what closes the citation gap.

The measurement problem this creates — and how to solve it now

Most marketing teams are measuring AI impact as a single aggregate number, if they're measuring it at all. That number hides more than it reveals.

Your bot traffic audit gives you a leading indicator most teams aren't tracking: AI crawler visit frequency by page. This is a proxy for AI engine interest. Pages with high AI crawler traffic are being used to synthesize answers about you. Pages with none aren't shaping what AI says about your brand, regardless of their SEO performance.

Pair that with the buyer prompt audit — run against a fixed set of 20–30 queries, tracked monthly — and you have a repeatable measurement loop. Citation frequency and citation accuracy by platform are the metrics that actually reflect what's happening in your buyers' research process. They are the upstream predictor of the pipeline that doesn't show up until someone books a call.

Forrester's analysis documents B2B traffic declines of 10–40% as buyers move research into AI. The companies closing that gap aren't waiting for their analytics stack to catch up. They're pulling data from sources they already have — server logs, CDN traffic reports, security tools — and reading what AI engines are actually doing with it.

The buyers are already there. The data is already there. The question is whether your team is looking at it.

Why the infrastructure behind this matters

The mechanism here is worth understanding clearly, because it's what makes any fix actually stick.

When a buyer asks ChatGPT or Perplexity about vendors in your category, those systems don't consult your website directly — they synthesize from what they've already been trained on or can retrieve in real time. What they retrieve and trust comes from publications with established editorial credibility: Forbes, TechCrunch, Harvard Business Review, Axios, the trade publications in your specific vertical.

Getting into those publications isn't a brand awareness play. It's infrastructure. It's how you build the earned authority that gives AI engines a credible, independent source to pull from when they're composing the answer a buyer just asked about your category. The bot traffic audit shows you where you stand. Earned media is how you close the gap.

This is what Machine Relations describes as the foundation of AI-era brand visibility: earned media in trusted publications creates the signal that AI engines treat as authoritative. Your owned content, even when it's technically optimized, is self-assertion. Third-party editorial coverage from recognized outlets is independent corroboration — the signal AI engines are built to prioritize.

The buyer research audit and the earned media strategy aren't separate projects. The audit tells you exactly which queries matter and which publications your category buyers trust. The earned media strategy fills those slots systematically. Together, they bring your brand into the AI answers your buyers are already reading, in the research session that happens before they ever contact sales.

If you want to see where your brand currently stands in AI answers, run the visibility audit here.

Your Buyers Are Already Researching You in AI. The Proof Is in Your Bot Traffic.

What Forrester's data actually shows

The three-step audit to find and use this data

The measurement problem this creates — and how to solve it now

Why the infrastructure behind this matters

Related Reading