How to Get AI Chatbot Traffic from ChatGPT, Claude, and Perplexity

AI chatbots collectively sent 3.5 million visitors to websites in March 2026 — and those visitors convert at 23 times the rate of organic search traffic, according to Ahrefs. That’s a tiny slice of total web traffic right now, but it behaves more like a referral from a trusted colleague than a click from a search results page. The marketers who understand this channel while it’s still uncrowded will have a structural advantage as AI search usage compounds in the months ahead.

What Happened

On May 15, 2026, Ahrefs published one of the most data-dense analyses of AI chatbot referral traffic to date, drawing on behavior tracked across 74,752 websites via chatgpt-vs-google.com. The findings challenge assumptions about where AI traffic fits in the marketing funnel and how to measure it accurately.

The headline number: all AI chatbots combined sent approximately 3.5 million website visitors in March 2026. ChatGPT dominated with 2.7 million of those visits — roughly 77% of all AI-driven clicks in the dataset. Perplexity and Gemini each contributed around 230,000 visitors, while Claude sent 102,000. For comparison, Google sent 345.2 million visitors in the same period, representing a 28.12% share of total web traffic. That puts AI chatbot traffic at just 0.28% of the total web.

On volume alone, that number looks ignorable. The conversion data is where the story changes fundamentally.

At Ahrefs, AI search visitors accounted for just 0.5% of total visitors but drove 12.1% of new signups — a 23x higher conversion rate than organic search. Buffer’s internal data, also cited in the Ahrefs piece, found AI referral traffic converting at 185% higher rates than organic: a 20.15% conversion rate for AI-referred visitors versus 7.06% for standard organic. These are not marginal improvements — they are category differences in traffic quality.

The growth trajectory compounds the case for paying attention now. Claude is growing fastest, at an average of 30.7% month-over-month, with a single-month spike of 153.5% in March 2026. Gemini is averaging 12.8% monthly growth. ChatGPT and Perplexity, as the more mature platforms, are growing more slowly at 1.4% and 2.9% per month respectively. A 30.7% monthly compounding rate puts Claude on a trajectory to become a material traffic source within 12 months, even starting from a smaller base.

The Ahrefs research also broke down which content formats earn AI citations most reliably. “Best” roundup posts dominate, comprising 43.8% of all page types that AI systems pull from. How-to guides, “Top” lists, and comparison posts (“X vs Y”) round out the top performers. The average AI prompt is 42 words long — and 75% of prompts are structured as commands (“best way to track,” “how to create”) rather than open-ended questions. This tells you a great deal about where in the buyer journey these interactions are happening.

There is also a citation-rate discrepancy that has direct implications for tracking and attribution. Even when Ahrefs is demonstrably a source for an AI-generated answer, the brand appears as a clickable link only 10–51% of the time across platforms. Perplexity leads with a 51.6% citation rate. ChatGPT and others frequently use content to generate answers without surfacing a trackable attribution — which creates a traffic attribution gap that most marketing teams have not yet accounted for.

The self-attribution data from Ahrefs’ own signup flow adds another layer. When users self-reported how they found Ahrefs in March 2026, Claude accounted for 2,836 sign-ups, ChatGPT for 1,978, Gemini for 619, and Perplexity for 67. The Claude number is striking given that Claude sent only 102,000 referred visitors — it implies an extraordinary conversion rate from that specific platform’s audience, possibly because Claude’s users skew toward professional and technical buyers.

Why This Matters

The 23x conversion differential is not an accident of the dataset. It reflects a fundamental difference in where AI-chatbot users sit in the buying journey compared to someone clicking a standard search result.

When someone types a 42-word command into ChatGPT or Perplexity — “what is the best project management software for a remote team of 50 that integrates with Slack and has a free tier” — they have already moved past the awareness stage. They are in active evaluation mode, often comparison-shopping or finalizing a shortlist. By the time they click through to your site from an AI citation, they have received a recommendation from a system they trust that specifically named you. You are not one of ten blue links competing for attention; you are a named recommendation from a system they have placed trust in.

Kevin Indig’s study of 48 high-stakes purchase decision participants — cited in the Ahrefs analysis — makes the psychology concrete. Of AI Mode users in the study, 64% clicked nothing at all during their research session. But 23% of those who did click did so specifically to confirm choices they had already made through the AI interface. The buyer journey is partially happening inside the AI platform before a single click occurs. By the time a visit registers in your analytics, the prospect may already be sold. That makes AI-referred visitors extraordinarily warm leads — which fully explains the conversion differential.

For agencies, this creates both an opportunity and a reporting accountability gap. The opportunity: clients who appear in AI citations in their category have a meaningful competitive advantage that compounds over time. The accountability gap: if you are not tracking AI referrals as a distinct channel and including them in client reporting, you are invisible to your clients in a channel that is delivering their best-converting traffic. Agencies that miss this will lose clients to competitors who can demonstrate they are actively managing it.

For in-house marketing teams at SaaS companies, the content strategy implications are significant and require adjustment. Google rewards pages with high topical authority, strong backlinks, and keyword density. AI systems reward content that is structured, fresh, editorially credible, and written in a format that a language model can extract a clean answer from. These are related but not identical optimization targets. Teams running a pure traditional SEO playbook may find their AI citation rates are materially lower than their search rankings would predict — because they are optimizing for a different retrieval mechanism.

For e-commerce and DTC brands, the earned media finding is the most actionable piece of the research. Semrush’s analysis cites research from Leoprd finding that 61.9% of brand citations within AI systems derive from editorial coverage, awards, and third-party reviews — not brand-owned content. If your product is not appearing in “best of” roundups on third-party sites, and if you are not winning industry awards, your product page is largely invisible to AI recommendation engines regardless of how well-optimized it is for search.

For solopreneurs and content creators, the channel is still remarkably uncrowded. Most of the 74,752 sites tracked by chatgpt-vs-google.com are not actively building AI citation strategies. The barrier to entry is lower than in traditional search because pure domain authority matters less than content structure and editorial credibility. A single well-structured, authoritatively sourced “Best X” post on a newer domain can outperform an established competitor’s AI citation rate.

Finally, there is the macro Google displacement story. According to chatgpt-vs-google.com, Google’s share of web traffic dropped from 35.11% in June 2025 to 28.12% in March 2026 — a nearly 7-percentage-point erosion in just 10 months. That is structural displacement, not seasonal noise. Semrush projects that AI search visitors will surpass traditional search visitors by 2028. The marketers who build their AI citation footprint in 2026 will be positioned well when that crossover hits.

The Data

The numbers from the Ahrefs analysis, chatgpt-vs-google.com, and Semrush tell a multi-layered story. The tables below pull the key data points into comparable formats for direct reference.

AI Chatbot Traffic Sources — March 2026

Source: Ahrefs, chatgpt-vs-google.com

Claude’s outsized self-reported sign-up number relative to its visitor volume (2,836 sign-ups from 102,000 visits) suggests a conversion rate that significantly exceeds ChatGPT’s ratio even though ChatGPT sends 26 times more traffic. This is why self-attribution data alongside referral tracking is essential — referral headers alone would have dramatically underrepresented Claude’s actual impact on the business.

Conversion Rate: AI Referral Traffic vs. Organic Search

Source: Ahrefs

Content Format Performance — AI Citation Distribution

Source: Ahrefs

Platform Citation Rates — Ahrefs Brand as Test Case

Source: Ahrefs

The citation rate gap explains why AI traffic is simultaneously underreported and undervalued. Platforms that frequently omit clickable source links drive what is known as “dark traffic” — sessions that arrive in GA4 as direct visits because no referrer header was transmitted. Per Semrush, actual AI-driven traffic volumes are “substantially underreported” in standard analytics tools. The 0.28% figure almost certainly understates the real number.

Content freshness is a statistically significant variable as well. Ahrefs found that AI assistants cite content that is 25.7% fresher than what organic search ranks, with a 13.1% preference specifically for recently updated pages. This is distinct from the freshness signals Google uses and is a direct attribute of how AI retrieval systems are weighted when selecting which content to surface in a response.

Real-World Use Cases

The data points to specific content, PR, and measurement patterns. Here are five concrete implementation scenarios across different marketing contexts, each built on what the research actually shows.

Use Case 1: SaaS Company Building “Best X” Content to Win Category Recommendations

Scenario: A B2B project management SaaS is invisible in AI recommendations when users ask ChatGPT “best project management software for remote teams of 50+.” Their product pages are properly indexed by Google, but they have no structured “best of” content that AI systems can extract clean answers from.

Implementation: Create a dedicated “Best Project Management Software for Remote Teams” post that covers the target use case explicitly in the H1 and key H2 headers. Include a comparison table with 8–10 competitors and explicitly state which tool is “best for” each specific use case alongside your own. Add a direct-answer summary paragraph in the first 100 words — a clean, extractable recommendation that AI systems can surface without guessing. Timestamp the page visibly (“Last updated May 2026”) and commit to monthly data updates. Include original data from customer surveys where possible. Submit the URL to Bing Webmaster Tools, which feeds OpenAI’s browsing index, and keep your XML sitemap current so new pages are discovered quickly.

Expected Outcome: “Best X” roundup posts represent 43.8% of all page types cited by AI systems per Ahrefs. A well-structured, regularly refreshed post in this format has the highest category-level probability of appearing in AI recommendations. Traffic volume will be modest initially, but the 23x conversion differential means even 200–300 AI-referred visits per month can generate more qualified pipeline than thousands of standard organic pageviews.

Use Case 2: Digital Agency Launching an AI Visibility Audit Service Line

Scenario: A performance marketing agency wants to expand its retainer menu beyond paid media and traditional SEO by offering an “AI Visibility Audit” — a monthly deliverable that tracks client brand appearance across ChatGPT, Perplexity, Claude, and Gemini.

Implementation: Build a standardized prompt library of 20–30 high-intent queries per client category (e.g., “best CRM for mid-market B2B companies,” “CRM software alternatives to Salesforce”). Run these queries manually or via API against each major AI platform monthly. Record brand mention rate, citation rate (whether a clickable link appears), competitor mention frequency, and the specific framing used to describe the brand when it does appear. Separately, create a GA4 custom segment filtering sessions from known AI referrer domains: chatgpt.com, claude.ai, perplexity.ai, and gemini.google.com. Add an optional self-reported acquisition field to client lead capture forms with AI chatbots listed as explicit options — this captures conversion volume that referral headers will miss. Deliver a monthly report comparing AI visibility scores against competitors and tracking improvement over time.

Expected Outcome: Agencies offering this deliverable are building a reporting capability that will be standard practice within 18 months. Semrush projects AI search traffic to surpass traditional search by 2028, meaning CMOs will increasingly require AI visibility measurement as a board-level KPI. Agencies that build the methodology and delivery infrastructure now — before it is commoditized — position themselves as AEO (Answer Engine Optimization) specialists with premium retainer positioning.

Use Case 3: E-Commerce Brand Leveraging Third-Party Editorial for AI Citations

Scenario: A DTC skincare brand wants to appear when users ask Claude or Perplexity “best moisturizer for combination skin under $50.” Their owned product pages are optimized for search, but they receive no AI citations for category queries.

Implementation: The core insight from Leoprd’s research (cited by Semrush) is that 61.9% of brand citations within AI systems derive from editorial coverage, not brand-owned content. Shift a meaningful portion of the content budget toward earned media: pitch products to beauty editors running “best of” roundups on high-authority publications. Submit to relevant beauty industry award programs. Encourage detailed customer reviews on third-party platforms using the specific language patterns of common AI queries. On owned product pages, add an above-the-fold summary structured as a direct answer to the query: “This moisturizer is formulated for combination skin types that need hydration without clogging pores.” Add FAQ structured data addressing the exact phrasing of the 5–10 most common AI queries your product should appear in.

Expected Outcome: Branded web mentions are the strongest correlating signal for AI answer appearance — Spearman correlation of 0.664 per Ahrefs. Each editorial mention compounds AI visibility for both branded queries and category queries. A 6-month earned media program targeting 15–20 relevant editorial placements should produce measurable increases in AI citation rates that are trackable through Ahrefs Brand Radar or systematic manual query monitoring across platforms.

Use Case 4: B2B Content Team Running a Systematic Freshness-First Refresh Program

Scenario: A marketing technology company has a library of 200+ blog posts from 2022–2024. Several rank on page 1 in Google for valuable keywords, but the team receives no meaningful AI referral traffic and suspects its static, aging content is being passed over by AI retrieval systems in favor of fresher material.

Implementation: Audit the post library for articles covering high-intent AI query topics: “best email marketing software,” “marketing automation tools comparison,” “how to set up lead scoring.” Prioritize 15–20 posts for a dedicated refresh cycle. For each: update all statistics to current data, rebuild the comparison table with current market options, add an explicit “Best for:” summary section in the first 100 words of the post, add FAQ structured data via schema markup answering the current phrasing of relevant queries, refresh internal links to point to current product pages, and update the publication date to reflect the actual refresh date. AI assistants show a 13.1% preference for recently updated pages and cite content that is 25.7% fresher on average than what organic search ranks per Ahrefs. HubSpot’s refresh of a single small business ideas post generated 1,135 new AI Overview mentions — a proof of concept for the leverage available through systematic updating.

Expected Outcome: A quarterly refresh cycle targeting 15–20 priority posts can generate measurable AI citation gains within 30–60 days of each republication. The compound effect over 12 months — running three to four cycles — positions the content library as a consistently fresh, AI-citation-ready asset. Track results by monitoring GA4 AI referral segments before and after each refresh wave, and run systematic prompt testing across major AI platforms to directly confirm citation gains.

Use Case 5: Consultant Building AI Visibility for Personal Brand Lead Generation

Scenario: An independent B2B marketing consultant wants their name or positioning to appear when potential clients ask Claude “best marketing consultant for B2B SaaS companies.” Currently they are invisible in AI responses to category queries, despite strong Google rankings for some long-tail keywords.

Implementation: The editorial credibility signal is the primary lever — not owned content. Publish detailed case studies on your own site with quantified outcomes (“increased inbound MQLs by 47% in six months for a Series B SaaS company”). Get featured in credible roundups: “best B2B marketing consultants” or “top growth consultants for SaaS” lists on authoritative industry publications. Guest post on marketing publications with strong editorial credibility. Seek speaker spots at recognized industry conferences — each generates editorial coverage that AI systems treat as an expertise validation signal. On your own site, create a services page that directly answers the query in plain language: “I work with B2B SaaS companies at $5M–$50M ARR that need to rebuild their demand generation function.” Add Person schema structured data explicitly categorizing your area of expertise. Target Perplexity monitoring first — at a 51.6% citation rate, Perplexity is the platform most likely to surface a clickable link when it uses your content.

Expected Outcome: Consultants who invest 6 months in systematic PR, editorial placement, and structured on-site content should see measurable improvements in branded and category AI citation rates. Given that Claude generated 2,836 sign-ups for Ahrefs in a single month with only 102,000 referred visits, even small volumes of AI-referred inbound inquiries can represent meaningful pipeline for a solo practice operating with high average contract values.

The Bigger Picture

The Ahrefs data lands at a genuinely inflection-point moment in how information flows across the web. Google’s nearly 7-percentage-point erosion in web traffic share between June 2025 and March 2026 is not gradual drift — it is structural displacement happening faster than most analysts projected.

Semrush projects that AI search traffic will surpass traditional search traffic by 2028. Given Claude’s 30.7% average monthly growth and Gemini’s 12.8%, that timeline may be optimistic for AI’s growth rather than pessimistic — those growth rates, if even partially sustained, compress the timeline toward 2027 rather than 2028. The platforms growing at 30% monthly are not operating on the same trajectory as those growing at 1–2%.

What is changing beneath the surface is the architecture of trust and authority through which web traffic flows. Traditional SEO built around link equity and domain authority — a system that took years and significant resources to accumulate, and that significantly advantaged incumbents. AI citation is building around editorial credibility and content structure — a system where a single well-executed, authoritatively sourced, properly formatted post can outperform years of SEO investment. This is simultaneously democratizing for newer publishers and disruptive for established ones whose advantage was built on link graphs.

The Leoprd data point (cited by Semrush) that 61.9% of AI brand citations come from third-party editorial coverage signals a structural rehabilitation of PR as a growth lever in a way that has not been true since the early days of link-building. For most of the past decade, PR’s relationship to SEO was mediated entirely by backlinks — coverage mattered because links from editorial outlets boosted domain authority scores. AI citation changes the relationship fundamentally: editorial coverage now matters because AI systems use it as a direct trust signal, independent of link equity. Brand mentions without links, industry award wins, analyst citations, conference keynotes — these are now ranking factors in a way they never were for Google’s algorithm.

The “dark traffic” problem documented by Semrush also has systemic implications for the entire industry’s understanding of this channel. If AI platforms routinely strip referrer headers — causing AI-driven sessions to register as direct traffic in GA4 — then the true scale of AI’s influence on web traffic is already substantially larger than any published 0.28% figure captures. Marketing teams that have seen unexplained lifts in direct traffic conversion rates over the past 12 months may already be benefiting from AI referrals without the measurement infrastructure to confirm it.

The April 2026 pullback tracked by chatgpt-vs-google.com — where both Google and ChatGPT saw traffic declines of roughly 10–12% month-over-month, with Gemini as the outlier at just -0.4% — is an important reminder that this channel is still volatile and subject to algorithmic and behavioral fluctuations. Building for AI citation requires patience and a multi-quarter view alongside tactical execution. Short-term month-over-month tracking will generate noisy signals; what matters is the 6–12 month trend.

What Smart Marketers Should Do Now

These five action items are sequenced by impact and immediacy. Each can be started this week without significant new tools or additional budget.

1. Install proper AI referral tracking in GA4 before doing anything else. You cannot optimize what you cannot measure, and right now most teams are flying blind on this channel. Create a GA4 custom segment capturing traffic from known AI referrer domains: chatgpt.com, claude.ai, perplexity.ai, gemini.google.com, and bing.com (which feeds Microsoft Copilot). This segment will undercount actual AI traffic because of the referrer-stripping problem, but it gives you a defensible baseline to improve against. Simultaneously, add a self-reported acquisition question to your signup or lead form with AI chatbots listed as explicit options. The Ahrefs self-attribution example is instructive: Claude drove 2,836 sign-ups in March 2026 that would have been invisible in standard referral tracking. Run both tracking methods in parallel for the most complete picture of what AI is actually generating.

2. Audit and build your “Best X” content inventory. Since “Best” roundup posts represent 43.8% of all page types pulled by AI systems, this is the single highest-ROI content format for AI citation and should be the first place you invest. Identify your 5–10 most valuable product categories, use cases, or audience segments. For each, ask: do we have a “best [X] for [specific use case]” page? If not, build one. If yes, audit it: Is there a comparison table? A direct-answer summary in the first 100 words? Current statistics (within 6 months)? A visible last-updated date? Inline citations to credible sources? Fix every gap before investing in other content formats. This is the foundation that everything else builds on.

3. Launch a quarterly content freshness program as a standing operational routine. AI systems cite content that is 25.7% fresher than what organic search ranks, with a 13.1% preference for recently updated pages per Ahrefs. Set a repeating quarterly calendar block to refresh your top 20 posts. The refresh protocol for each post: update every statistic to current data, rebuild comparison tables with current market options, add a new FAQ section targeting current query phrasing, and update the publication date to reflect the actual refresh. This is a signal reset that tells AI retrieval systems your content reflects current reality rather than a past snapshot. HubSpot’s single-post refresh generated 1,135 new AI Overview mentions — a systematic 20-post quarterly program compounds that leverage significantly over a year.

4. Shift part of your content budget toward earned media and editorial placements. The research finding that 61.9% of brand citations in AI responses derive from editorial coverage — not owned content — requires a real budget reallocation response, not a mental note. Identify the top 10–15 third-party publications in your category that publish “best of” roundups. Build a systematic outreach program to their editors. Pursue two to three relevant industry award applications per year. Seek analyst coverage from recognized firms in your space. Each of these generates the editorial mention signal that Ahrefs found correlates most strongly with AI answer appearance (Spearman 0.664). This is PR as an SEO lever in a way that has not been true in a decade.

5. Target Perplexity specifically as your primary AI citation monitoring platform. Perplexity has the highest citation rate of any major AI platform at 51.6% per Ahrefs — more than double the lower end of the range across platforms — meaning content that Perplexity draws from is dramatically more likely to generate an actual click than content used by lower-citation-rate platforms. Perplexity rewards well-sourced, data-rich content: posts with inline citations, original research, clear comparison tables, and structured direct-answer writing. Reformat your highest-value content for Perplexity’s retrieval patterns: put the direct answer in the opening paragraph, use descriptive H2s that mirror the phrasing of common queries, and include at least one data table per post. Run monthly manual checks querying Perplexity with your top 20 target queries to monitor citation rates and track competitor presence week over week.

What to Watch Next

Several specific developments will reshape the AI chatbot traffic landscape materially over the next 6–12 months and deserve active monitoring.

Claude’s sustained growth rate and distribution partnerships. At 30.7% average monthly growth and a 153.5% spike in March 2026, Claude is the most volatile variable in this space. Anthropic’s enterprise and consumer product distribution announcements are the leading indicator. Deeper integrations into productivity software — Slack, Microsoft 365, native browser toolbars — would accelerate Claude’s referral traffic volume significantly and push self-reported attribution from Claude even higher. Watch Anthropic’s partnership announcements through Q3 2026 as a signal for where Claude’s traffic trajectory is heading.

Google AI Mode expansion and non-click behavior. Kevin Indig’s finding that 64% of AI Mode users clicked nothing during high-stakes purchase research is a critical data point for all publishers to track. As Google expands AI Mode to more query types and geographies through Q2 and Q3 2026, the non-click pattern may accelerate Google’s traffic erosion even beyond the 7 percentage points already lost. AI Mode is potentially training a generation of users who never develop the habit of clicking through — which makes winning AI citations on external platforms even more strategically important.

Referrer header standardization. Industry pressure is building for AI platforms to consistently transmit referrer information so publishers and marketers can accurately attribute traffic. If OpenAI, Anthropic, and Google standardize referrer headers across their chatbot interfaces — whether through regulatory pressure or voluntary publisher agreements — the “true” volume of AI-driven traffic will become visible in GA4 for the first time. Watch for IAB Tech Lab or publisher coalition announcements on this issue in the second half of 2026, as it would trigger a significant upward revision in reported AI traffic share.

Platform content licensing deals. Several major publishers have already signed content licensing agreements with AI companies. How these deals affect organic citation patterns — specifically whether licensed content receives algorithmic preference in AI answers — is something to monitor carefully in Q3 and Q4 2026. If licensed content does receive preferential placement, the playbook for independent publishers and brands changes substantially, and the importance of earned third-party editorial coverage increases even further.

Bottom Line

AI chatbot traffic represents just 0.28% of total web traffic today and converts at 23 times the rate of organic search. The volume is small; the quality is exceptional; and the growth trajectory — Claude at 30.7% average monthly growth, Gemini at 12.8% — means the gap between AI and traditional search traffic is closing faster than most marketing plans currently account for. The tactics required to earn AI citations are not radically foreign: structured “Best X” content, a systematic freshness program, earned editorial coverage, and credible sourcing are all established disciplines. What is different is the deliberate restructuring of that work for AI retrieval patterns and the measurement infrastructure required to see what it’s actually generating. The marketers who build their AI citation footprint now, while the channel is still uncrowded and the Semrush 2028 crossover projection still feels distant, will be the ones best positioned when AI search traffic becomes the primary channel rather than a footnote.