Sew What’s Next AEO Experiment
A measurable, public test of whether a new niche product can earn visibility in AI-generated answers.
Experiment Overview
A public experiment tracking whether a new sewing app can become discoverable in AI-generated answers for sewing project planning, fabric stash organization, and pattern-matching queries.
The goal is not only to improve the app’s visibility, but to document a repeatable AEO process: establish a baseline, analyze competing answers, make targeted content and technical changes, rerun the test, and publish what changed.
Methodology
- Establish a baseline. Run a fixed set of 15 prompts and record mentions, citations, competitors, and answer patterns.
- Analyze competing answers. Study the products and sources surfaced for sewing planning, fabric stash, and pattern-matching queries.
- Make targeted changes. Improve first-party content and technical signals based on gaps revealed by the baseline.
- Rerun and document. Repeat the frozen prompt set, compare results, and publish what changed.
Baseline
The first frozen test establishes the starting point. These numbers do not claim improvement; they give future experiment rounds something consistent to compare against.
- Prompt set
- 15 prompts per platform
- Total responses
- 60
- Brand mentions
- 0/60
- Citation rate
- 0%
Build the AEO Foundation
After establishing a 0% AI visibility baseline, I focused on making Sew What’s Next easier for search and answer engines to discover, understand, and connect with its core use case.
What I implemented after the baseline
Launched dedicated domain sew-whats-next.com.
Created homepage that clearly described Sew What’s Next as a sewing planning app that matches fabric you already own with sewing patterns.
Published a supporting page targeting the prompts being tracked.
Added real product snapshots to demonstrate a functioning app.
Added technical schema and fixed the sitemap to match the new domain.
Verified crawlability, submitted the page and sitemap through Google Search Console, and began monitoring indexing.
ChatGPT visibility reached 20%
After Phase 1, Sew What’s Next appeared in 3 of 15 ChatGPT responses. All three recommendations cited sew-whats-next.com.
ChatGPT visibility increased from 0 percent to 20 percent. Claude, Google AI Mode, and Gemini remained at 0 percent.
- 3 of 15cited recommendations
- 2ranked first
- 1ranked second
A note about the result: The increase happened after Phase 1, but this test cannot prove that those changes caused it. AI answers can change from one test to the next.
What I’ve learned
Clear positioning helped ChatGPT find and cite the product. After I launched the website and added copy focused on what I wanted the app to be known for, ChatGPT surfaced it for specific prompts in a niche without a clear category leader.
Publishing a website does not make it a trusted source. This matches my SEO experience: “If you build it, they will come” is not true. Google and Gemini did not cite the site simply because it was online. A new site still needs trust signals such as backlinks, external references, and domain authority.
Other answer engines still favored established sources. Claude, Gemini, and Google AI Mode surfaced better-known products that have been around longer and appear in “best sewing app” articles. This matches AEO research I’ve read suggesting that Claude gives significant weight to third-party mentions from established media outlets.
Results over time
- Round 0 · Baseline
ChatGPT visibility began at 0 of 15 prompts, or 0%.
- Measurement 1 · After Phase 1
ChatGPT visibility reached 3 of 15 prompts, or 20%. Claude, Google AI Mode, and Gemini remained at 0%.
Next: Build authority beyond the website
Phase 2 will test the next set of discoverability signals, then repeat the same prompt set to measure what changes.