/ai-search1,088 words

Citations Are Not Enough: Conductor VP Names the 'Golden Metric' of AI Search

Conductor's Patrick Reinhart says the golden AEO metric is whether an LLM recommends your brand — not citations, mentions, or traffic. YouTube tops cited sources in Conductor's data.

Patrick Reinhart On The Golden Metric In AI Search: Brand Recommendations via @sejournal, @lorenbaker
Patrick Reinhart On The Golden Metric In AI Search: Brand Recommendations via @sejournal, @lorenbakerAI-generated
  • YouTube is the number one most-cited source in Conductor's data; seven of the top 15% of cited sites are social platforms.
  • About 90% of Conductor's customers are large enterprise organizations, concentrated in finance, insurance, and healthcare.
  • Reinhart calls LLM recommendation quality the 'golden metric' and says no vendor yet tracks it at scale.
  • Conductor outranked a third-party site carrying outdated 2018 product claims within days by creating counter-content.
  • Reddit's citation visibility in GPT dropped sharply roughly a month before the podcast recording, per Loren Baker.

YouTube is the number one most-cited source in Conductor's AI visibility data, and seven of the top 15% of cited sites are social platforms. That finding anchors Patrick Reinhart's core argument: in answer engine optimization, being recommended by an LLM matters more than being cited, and both matter more than ranking.

Reinhart, VP of Services and Thought Leadership at Conductor, laid out his framework on the Search Engine Journal Podcast with founder Loren Baker. He has spent 11 years at Conductor, and four or five years as a customer before that. Conductor serves an enterprise base: about 90% of its customers are large organizations, concentrated in finance, insurance, and healthcare, with ecommerce and grocery brands such as H-E-B and BJ's also on the roster.

What is the 'golden metric' in AI search?

"Are you being recommended by an LLM? That's the golden metric," Reinhart said.

Citation rate is currently the leading AEO metric. Brand mentions matter too, but Reinhart treats them as directional. Not all mentions are equal — ChatGPT can say "Do not work with that company because of XYZ," a negative mention he has seen multiple times.

The failure modes require different fixes:

  • No mention at all: a content gap problem.
  • Negative mention: a third-party conversation problem.
  • Mention without recommendation: a persuasiveness problem.

Reinhart said no vendor currently tracks recommendation quality from a technology standpoint. Most tools offer visibility scores that average several signals. Conductor is building tooling for the golden metric, scoring whether each appearance was a positive recommendation or a negative non-recommendation.

How do enterprises actually approach AEO today?

Most AEO strategies remain heavily grounded in traditional SEO, Reinhart said, but measured and thought about very differently because LLMs behave differently from search engines. Enterprise teams are upskilling internally rather than hiring. He predicts a hiring "gold rush" in roughly a year, when programs built by people inside large organizations now come to fruition.

He stressed LLM behavioral differences. Claude behaves vastly differently from Gemini. Most market share sits with ChatGPT and Gemini; people use Claude to build, but use ChatGPT for everything.

Practitioners must build a deterministic set of prompts that accurately represents how a brand appears across LLMs, since tracking every prompt variant is impossible. Data fidelity varies widely: some vendors buy panel data, Conductor uses direct APIs, and scraping Google keeps getting harder — a point underscored by a rough week for rank-tracking tools.

What happens when AI gets your brand wrong?

Baker raised a concrete scenario: Reddit's citation visibility in GPT dropped sharply about a month ago, yet GPT appears to still use thread content without surfacing Reddit as a source. Outdated facts — a two-year-old thread with wrong warranty pricing — can still poison answers.

Reinhart said fact-checking will become a make-or-break vendor feature. Conductor's own experience: outgoing CEO Seth Besmertnik found a random site carrying 2018 information claiming Conductor lacked one of its core features. The site would not update, so Conductor created counter-content and outranked it within days.

Users can already connect to Conductor's MCP in Claude, send snippets plus brand guidelines, and ask whether the information is correct. Constant automated crawling for outdated third-party mentions is the next step, with major innovation expected over the next 12 months.

"When I talk to leaders, they don't ask, 'Where do I show up?' They ask, 'How are people talking about me? Is it accurate?'" Reinhart said.

Where do SEO and AEO diverge?

The tactics overlap: fill content gaps, build publisher partnerships, produce YouTube videos. The biggest divergence is measurement. Traditional metrics — rankings, backlinks, traffic — have become directional.

"Don't set your goals around traffic anymore. That's a fool's errand. It's too volatile right now," Reinhart said.

The new measurement stack includes:

  • Positive versus negative citations and recommendations.
  • Sentiment as a standing concern, not a crisis response.
  • Log file analysis to see whether LLMs fetch and serve your content, and to detect "dead zones" they stop returning to.

AEO is a team sport, Reinhart argued. Search data must flow to social, PR, and executive teams. He cites Rand Fishkin's observation that third-party content suddenly matters because of AI — while noting "search everywhere optimization" always existed.

Why does the C-suite care now?

The pressure flows top-down for the first time. Reinhart has talked to more C-level executives in the last six months than in the previous six years. The topic has reached the mainstream — his brother, a banker, called asking about AI after reading The Wall Street Journal.

Baker shared an illustrative anecdote: a client's Shopify relaunch never set canonical tags. His tickets stalled for months until AI Overviews visibility dropped from URL cycling. He filed a new ticket labeled "911 emergency AI issue." The fix landed in two days.

How do brands earn AI citations?

AEO is a brand authority exercise, not a domain authority exercise, Reinhart said. Reddit's citation drop was inevitable as brands produced more content, but Reddit remains an authenticity signal in the consideration phase even when uncited.

In Conductor's data, YouTube ranks first among cited sources, and seven of the top 15% of sites are social platforms — Facebook, LinkedIn, YouTube, and Reddit among them. LLMs seek real people having real conversations, in Reinhart's framing. He describes search engines as living in a 2D index while LLMs inhabit a 3D world, drawing from communities like Quora and YouTube.

Named experts are a major lever. Reinhart criticized the many sites still lacking authors, author schema, and article schema. LLMs can credentialize individuals on LinkedIn based on career longevity and industry contribution — something traditional search engines could not readily do.

Where should brands invest for 2027?

Reinhart's three priorities:

  1. Invest in expertise. Identify people in the organization, tie content to them, build author pages, and use author and article schema.
  2. Make more video. Start a podcast, join webinars, and accept early imperfection.
  3. Fix the process. Build a flywheel so social, PR, and content teams act on search signals without a six-month internal battle.

Process is the most important of the three, he said, because speed compounds when AI throws surprises at the program. The tactics may hold steady even as the landscape shifts within six months.

For monitoring, watch whether vendors ship recommendation-level tracking and automated fact-checking — the two features Reinhart calls necessary for any AI visibility technology to stay viable in the coming years.

via linkedin.com (Original)

More from Nathan Brooks

Nathan Brooks

Show full bio

Senior reporter covering companies, deals and strategy in Search rankings and algorithm updates, generative engine optimization at SERP Journal.

226 articles