As conversational search engines and artificial intelligence assistants steadily reshape how consumers discover local businesses, marketers face a daunting new challenge: measuring the unmeasurable.
During a high-impact session at SEJ Live, Sean McCrohan, Vice President of Technology at CallRail, and Steve Wiideman, an enterprise SEO consultant advising multi-location brands, sat down to unpack the shifting dynamics of AI search traffic. The data they presented reveals a complex ecosystem where AI-driven referrals are quietly climbing, yet traditional tracking mechanics fall short.
While AI citation clicks still account for a modest slice of total volume, their rapid growth—coupled with unique temporal trends and tracking hurdles—demands an immediate evolution in how multi-location and local businesses approach digital marketing, analytics, and conversion rate optimization.
Main Facts: The State of AI Visibility and Local Search
The core takeaway from the SEJ Live session is that while AI engines like ChatGPT, Google Gemini, Claude, Perplexity, and Grok are increasingly influencing consumer decisions, capturing and measuring that impact requires a fundamental shift in strategy.
- Modest Yet Growing Share: AI citation clicks currently account for roughly 1% to 2% of total calls for CallRail’s customer base. Though seemingly small, this metric has roughly doubled since January, signaling an upward trajectory that mimics the early days of organic search and mobile discovery.
- The Attribution Black Hole: Unlike traditional paid search or conventional organic traffic, tracking AI activity all the way from an initial prompt to a phone call remains exceptionally difficult. Ad networks from AI vendors are still in their infancy compared to mature ecosystems like Google Ads, leaving marketers to rely heavily on self-reported attribution and call recordings.
- The 24/7 Shift: AI traffic behaves differently than traditional web traffic. Historically, about half of a brand’s website traffic occurs outside regular business hours. With AI search, that figure climbs to nearly two-thirds, as consumers use conversational assistants deep into the night.
- Server-Side Tracking is Essential: Because AI web crawlers generally do not execute client-side scripts, standard in-browser phone number-swapping tools fail to capture AI-referred data. Marketers must pivot to server-side number swaps to properly attribute AI crawler behavior.
Chronology: How the Conversation Around AI Attribution Evolved
The discussion at SEJ Live highlighted a rapid evolution in how technical marketers view AI search engines—transitioning from novelty experiments to vital discovery channels.
January: The Baseline
At the start of the year, AI-driven citation clicks and referrals hovered near fractional percentages for most localized brands. At the time, conversational engines were largely viewed by enterprise brands as auxiliary tools rather than intentional local discovery platforms.
Spring and Summer: The Volatility Phase
As platforms like OpenAI’s ChatGPT and Google’s Gemini rolled out deeper browsing capabilities, marketers noticed wild fluctuations in how often brands were cited. Analyses, such as those from Steady Demand, revealed significant "prompt drift"—the phenomenon where repeating the exact same local search query yields overlapping sources only about 40% of the time, and identical top-ranking businesses only 7% of the time (a stark contrast to the 90%+ consistency seen in traditional Google local packs). Furthermore, abrupt shifts, such as Reddit’s sudden dip in AI citations during August, underscored the volatility of relying on single citation sources.
August 26: SEJ Live Revelation
During the second session of SEJ Live, Sean McCrohan revised an earlier presentation slide that had put AI-driven growth at 65% over six months. Updated tracking revealed that measurable AI referral metrics had actually more than doubled (rising over 100%) since January, cementing the realization that AI search is no longer a peripheral trend but a permanent, growing component of the customer journey.
Supporting Data and Metrics: What the Numbers Tell Us
Marketers navigating the AI optimization landscape must rely on specific, measurable indicators while accepting the current limitations of analytics platforms.
- The 1%–2% Threshold: Filtering Google Analytics 4 (GA4) traffic for specific AI user-agents across multi-location brands reveals that direct referrals hover around 1%. Wiideman characterizes this as entirely normal for the early maturity stage of a dedicated discovery channel.
- Audience Awareness: A live audience poll during the session revealed a surprising split: approximately 25% of attendees reported they could definitively tell when an AI assistant sent a lead, while nearly 50% remained unsure. McCrohan noted that the high percentage of "yes" responses exceeded his initial expectations.
- The Power of Reviews: Data cited during the event emphasizes that maintaining an average review score between 4.5 and 4.7 stars across platforms like Google Business Profile, Yelp, TripAdvisor, and Apple Maps is critical. Falling below this threshold dramatically reduces a business’s likelihood of being recommended by an AI agent.
- Prompt Library Scale: SEO strategies are shifting toward building semantic triples—specific, verifiable factual claims about a business. Experts recommend curating a prompt library of 100 to 125 core queries to monitor sentiment, visibility, and prompt drift over time.
Official Responses and Industry Perspectives
The perspectives shared by McCrohan and Wiideman illuminate the friction between technological capability and enterprise readiness.
Sean McCrohan emphasized the architectural hurdles of tracking AI traffic, noting that current AI vendors have been reluctant to establish the robust, granular ad networks marketers are accustomed to. Regarding technical implementation, McCrohan was blunt about web scraping behavior:
"The crawlers that AI agents are using to retrieve this data do not execute scripts on your page when they go to look at your website. They just don’t."
McCrohan also stressed the operational urgency for local businesses, pointing out that AI assistants often recommend the top three options, meaning a missed call instantly hands a lead to a competitor. "Get something, because more and more of this business is a 24-hour business," he advised.
Steve Wiideman focused heavily on governance, structure, and data integrity for multi-location brands. He warned against haphazard user-agent testing due to risks surrounding search engine spam and cloaking policies. Instead, Wiideman advocated for centralized corporate control over schema markup, data feeds, and NAP (Name, Address, Phone) consistency across the entire web ecosystem.
Implications: What Marketers Must Do Next
The rise of conversational search and AI assistants carries profound implications for digital marketing strategies, technical SEO, and lead generation frameworks.
1. Shift from Keywords to Natural Language and Transcripts
Because AI models synthesize human language, static keyword lists are no longer enough. McCrohan noted that analyzing customer call transcripts and site chat logs reveals the exact vernacular, phrasing, and pain points consumers use—phrases that rarely appear in sterile corporate copy. Aligning website content with these real-world conversations increases the likelihood of being cited by an LLM.
2. Upgrade to Server-Side Infrastructure
Marketers relying on traditional client-side JavaScript tags for dynamic phone number insertion or visitor tracking will remain blind to AI crawler activity. Enterprise brands must transition to server-side handling to ensure accurate attribution as privacy regulations tighten and AI crawler usage expands.
3. Embrace Cross-Platform Reputation Management
AI assistants do not pull data exclusively from a brand’s website. They aggregate insights from Reddit, Yelp, TripAdvisor, and niche industry forums. Multi-location brands can no longer afford to focus solely on Google Business Profile; active reputation management across all major review ecosystems is now a fundamental ranking and recommendation signal.
4. Enterprise Governance Over Local Chaos
For brands managing dozens or thousands of locations, rogue listings and unauthorized local phone numbers poison the data feeds that AI agents rely on. Centralized governance—where corporate teams strictly manage schema, citations, and structured data—is more valuable for AI visibility than traditional local ranking tricks.
Looking Ahead
As Steve Wiideman predicted, the near future will likely see AI agents independently verifying contact numbers and initiating direct calls on behalf of the consumer, bypassing the brand’s website entirely.
Yet, despite the technical disruption, the fundamental rules of local business success remain grounded. As McCrohan aptly summarized: "It is an exciting new world, but it is not a 100% new world. There is carryover. It will be OK."
