TL;DR

AI search engines decide who to cite before a user ever types a query. The sourcing pool is shaped by entity signals, third-party mentions, and original data — not by the content sitting on your website. Adam Tanguay's May 2026 analysis in Search Engine Land makes this explicit: traditional SEO influences the click, but pre-query foundation determines the citation. This article translates that insight into a measurable 30-60-90 day action plan.

Why "Publish and Pray" No Longer Works

Rand Fishkin's landmark March 2026 study, Influence Happens Everywhere, analyzed the 5,000 most-visited sites across mobile and desktop. The finding that rattled the industry: while Google still commands 73% of search traffic, search itself is merely a response to influence created elsewhere.

Tanguay builds on this with a critical observation. When ChatGPT, Perplexity, Claude, or Gemini assembles an answer, the model has already "decided" its sourcing pool based on what it learned during training and what its retrieval layer can find in trusted indices. Your blog post published last Tuesday is competing against months of accumulated entity signals and third-party mentions.

Dave Davies, principal SEO manager at Weights & Biases, reinforced the same thesis in his December 2025 piece for Search Engine Land: mentions, citations, and structured visibility signals are becoming the new levers of trust — and the path to revenue.

The Three Layers of Pre-Query Foundation

Tanguay's framework maps neatly onto three measurable layers. Each one feeds the AI citation pool independently.

1. Entity Layer

This is the structural bedrock. AI systems identify brands through entity recognition — Wikipedia entries, Crunchbase profiles, LinkedIn company pages, Google Knowledge Panels, and structured data (JSON-LD) on your own site. If an AI model cannot confidently identify what your brand is, it will not cite you regardless of how good your content is.

Audit checklist: Does your brand have a Wikipedia entry or Wikidata item? Is your Crunchbase profile current? Does your homepage carry Organization schema with sameAs links to all official profiles?

2. Mention Layer

This is where third-party validation lives. AI retrieval systems prioritize sources that are mentioned, cited, or discussed by other authoritative publishers. Think guest articles in industry publications, podcast appearances that generate show notes and backlinks, Reddit and Quora threads where real users recommend your product, and analyst reports that name your company.

The key insight from Tanguay: content on your own website is a response to influence, not the cause of it. Your owned content helps with traditional search. Third-party mentions feed the AI citation pool.

3. Original Data Layer

AI models crave citable data. Proprietary research, benchmark studies, survey results, and unique datasets give AI systems something they cannot synthesize from generic content. Davies' analysis of 7.2 million sessions confirmed that content earning sustained AI citations shared a consistent profile: original data, expert voice, and clear structure that an AI system could extract and attribute.

Critical mistake to avoid: gating your research behind a PDF download form. As Tanguay puts it, crawlable, citable original data earns AI attribution. A PDF behind a form wall earns nothing as discovery migrates to AI interfaces.

The 30-60-90 Day Action Plan

Days 1–30: Entity Audit and Cleanup

Query ChatGPT, Claude, Gemini, and Perplexity with the prompts your customers actually use. Note which brands appear and which sources get cited. If you are not among them, adding more content is not the first fix — fixing your entity signals is.

Deliverables: complete entity audit across all four major AI platforms, fix or create structured data on your homepage, claim and update Crunchbase and Wikidata entries, establish a baseline mention count.

Days 31–60: Five Third-Party Placements

Secure five meaningful third-party mentions. These could be guest posts in industry publications, podcast appearances, contributed data to analyst reports, or substantive answers in community forums where your expertise is relevant. Each placement should link back to your brand entity, not just your homepage.

Days 61–90: First Original Data Report

Publish one piece of original research that is crawlable, structured, and designed to be cited. Include clear methodology, concrete numbers, and a summary section that AI retrieval systems can extract cleanly. Make it available as both a web page and a markdown file for AI crawlers.

How to Measure Progress

The traditional SEO dashboard will not tell you whether your pre-query foundation is improving. You need to track: how often AI platforms mention your brand in response to relevant queries, which competitors appear in your place, and whether your entity signals are being recognized correctly.

Tools like Be Recommended let you monitor your mention count across ChatGPT, Perplexity, Claude, and Gemini in one dashboard — so you can see whether your 30-60-90 day work is actually shifting your position in the AI citation pool.

FAQ

Does traditional SEO still matter if AI visibility is pre-query? Yes. Traditional SEO drives direct search traffic, which remains the largest single traffic source for most websites. But AI citation — which influences purchase decisions increasingly — requires a separate, complementary strategy focused on entity signals and third-party mentions.

How long before I see results in AI citations? Typically 4–12 weeks after establishing strong entity signals and earning third-party mentions. AI models update their retrieval indices at different intervals, so results appear on Perplexity (near real-time) before ChatGPT (periodic index updates).

Can I influence AI citations without a large budget? Absolutely. Entity cleanup costs nothing. Community participation on Reddit and Quora is free. One well-structured original data report can outperform months of generic blog posts in AI citation impact.

Key Takeaways

AI visibility is determined before the query — by entity signals, mentions, and original data, not by your latest blog post.
Fishkin's 2026 data confirms search is a response to influence created elsewhere. Tanguay shows how to build that influence systematically.
The 30-60-90 framework gives marketing teams a measurable path: audit entities (30 days), earn mentions (60 days), publish original data (90 days).
Measure progress across all four major AI platforms, not just Google.

Sources

Adam Tanguay, Why AI visibility starts before search and ends with citations, Search Engine Land, May 2026
Rand Fishkin, Influence Happens Everywhere, SparkToro, March 2026
Dave Davies, Mentions, citations, and clicks: Your 2026 content strategy, Search Engine Land, December 2025

AI Visibility Starts Before the Search Query: A 30-60-90 Day Framework