Case Study
Project Type
AI Search
Industry
Real Estate

AI and Local Search for a Clarksville Realtor

Hodges Homes

We set out to get Chris Hodges found for local buyer and seller questions, in Google and in AI answers. Six weeks in, Google is moving and ChatGPT is not. Here is the honest picture.

The Client

Strong on his own name, invisible for local questions

Chris Hodges ranked near the top of Google for his own name, but nearly all of his search clicks came from people who already knew him.

In a September 2026 ChatGPT baseline, he was named on questions about himself but not on general questions about buying or selling in Clarksville.

Hodges Homes website homepage

What we built

Measure first, then publish answers to the questions local buyers and sellers ask.

Measurement
  • A 96-question ChatGPT baseline before any new content went live
  • Semrush rank tracking on 121 Google keywords and 63 ChatGPT keywords
  • A manual re-run of the 41 core ChatGPT questions in November
Content
  • FAQ library grown from 30 to 88 answers between September and October 2026
  • Three pillar pages added in October to test dedicated pages per topic
  • Structured data for the FAQs, the agent and the business, carried into the rebuild

Results

Semrush Google rank tracking (Clarksville, 121 keywords) and Search Console, from the September 28 baseline to November 10, 2026.

30 → 88
Published FAQ answers
6 → 8
Tracked keywords in Google's top 3
13 → 16
Tracked keywords ranking in Google's top 100
9 → 108
Search impressions for the FAQ page in four weeks, with its average position moving from 15.9 to 7.5

Read these as early signals

Two FAQs on assumable homes reached position 1 in Google. Search volumes are small, in the hundreds of impressions, and daily visibility swings week to week, so we read the trend rather than any single day.

Where This Stands

What has not moved yet

  • ChatGPT cites him on 0 of 41 questions. Our November re-run of the core local questions found Chris named and cited on none of them, the same as the baseline.
  • Standalone pages win the AI answers. The Clarksville sources ChatGPT cites are pages built around one question each. His answers sit together on one FAQ page. That is a hypothesis we are testing with the new pillar pages.
  • Agent recommendations follow review counts. For "who is a good realtor" questions, AI answers leaned on Zillow review counts, which a website alone cannot change.

Want to know how AI search sees your business?