All resources

How Graphite Increased AI Visibility from 0% to 33% in 6 Mentions

Delia Rowland

March 3, 2026

10

minutes read

Case study

The Challenge

When users ask AI assistants like ChatGPT or Claude for recommendations like, "What are the best SEO agencies?" the answers come from somewhere. But the question on every marketer’s mind is where.

Graphite, a leading SEO agency, wasn't showing up in those answers. Despite strong credentials and client results, they were invisible in AI-generated recommendations. For "best SEO agencies" queries specifically, they appeared in 0% of AI responses.

The question: Could that be changed?

The Hypothesis

LLMs don't invent recommendations. They pull from sources they've indexed, particularly third-party content like listicles, roundups, and comparison articles.

If Graphite could get mentioned in the right sources (the ones LLMs actually cite), they should appear more frequently in AI answers.

The Approach

Graphite partnered with Noble, a platform that automates brand mentions in third-party articles that LLMs tend to cite. Over three months (October–December 2025), Noble placed Graphite in 6 targeted articles:

Article Publisher Focus
Top Generative Engine Optimization (GEO) Agencies AvenueZ GEO
Best Answer Engine Optimization (AEO) Agencies Reverbico AEO
The 16 Best Enterprise SEO Agencies Omniscient Digital Enterprise SEO
Best AI SEO Agencies Digital Elevator AI/SEO
Top SEO Agencies San Francisco BabyLoveGrowth Local SEO
6 Best SaaS SEO Agencies RankTracker SaaS SEO

To measure impact, Graphite used Peec, a tool that monitors what percentage of AI responses mention a given brand for specific query sets.

The Results

Overall AI Visibility

Graphite's share-of-voice across all tracked SEO-related prompts:

  • Before: ~0.4%
  • After: ~1.7%
  • Change: 4.5x increase

High-Intent Queries

For "best SEO agencies" queries specifically:

  • Before: 0% (never recommended)
  • After: ~33% of answers
  • Change: From invisible to 1 in 3 AI recommendations

Correlation with Mentions

As each mention went live and was indexed by LLMs (estimated 3-day lag), visibility increased correspondingly. The correlation between "active indexed mentions" and daily visibility was 0.37 - a moderate positive relationship consistent with the hypothesis.

The Timeline

The data shows a clear pattern: visibility stepped up as mentions accumulated.

Early Period (Oct 3–12): Before any mentions were indexed

  • Average visibility: ~0.4%
  • "Best SEO agencies" mentions: 0%

Mid Period (Oct–Nov): First 3 mentions live

  • Visibility begins climbing
  • Early mentions showed strong local impact (one jumped from 0.1% to 0.8% in the week after indexing)

Late Period (Dec 26–Jan 1): All 6 mentions indexed

  • Average visibility: ~1.7%
  • "Best SEO agencies" mentions: ~33%

What This Suggests

The pattern is consistent with a straightforward mechanism: LLMs treat third-party roundups like a dynamic shortlist. Get mentioned in those sources, get recommended more often.

This isn't a controlled A/B test and there's no parallel universe where Graphite didn't get these mentions. But the timing, magnitude, and consistency of the uplift are hard to explain as coincidence:

  • Visibility was near-zero before mentions
  • Visibility increased as mentions went live
  • The "best SEO agencies" cluster went from 0% to 33% after relevant listicles were indexed
  • Correlation between active mentions and visibility was positive and steady

Key Takeaways

1. AI visibility is measurable. Tools like Peec make it possible to track share-of-voice in AI answers the same way you'd track search rankings.

2. Third-party mentions appear to matter. LLMs rely heavily on external sources for recommendations. Being present in the right articles correlates with appearing in AI answers.

3. The effect compounds. Early mentions lifted Graphite from near-zero. Later mentions reinforced and sustained the higher visibility rather than producing new spikes.

4. High-intent queries are winnable. Moving from 0% to 33% in "best SEO agencies" queries represents a significant competitive shift for a specific, high-value prompt cluster.

Methodology Notes

  • Measurement period: October 3, 2025 – January 1, 2026
  • Indexing assumption: 3-day lag from publication to LLM incorporation
  • Visibility metric: Daily share-of-voice (% of AI answers mentioning Graphite)
  • Limitations: Small sample (6 mentions), no control group, late mentions had truncated measurement windows (5–6 days post-index), pre/post windows overlapped

For the full technical methodology and raw data, contact Graphite or Noble.

Let’s win AI search

GET DISCOVERED ON