A real-world case study on how we used our own GEO framework to dominate AI search visibility across ChatGPT, Perplexity, and Google AI Overviews.
1. Context: Practicing What We Preach
By mid-2025, it became clear to the team at Proper Propaganda that AI search engines like ChatGPT, Google AI Overviews, and Perplexity had become discovery gateways for buyers of our services. So in July 2025, we turned our GEO framework inward. If we were going to help clients win in AI search, we needed to lead by example and show up ourselves. The leadership team outlined goals and set a firm commitment to bolster this piece of the business. This case study details what we did and what happened.
2. Defining the Battlefield
We began with a diagnostic phase modeled exactly on the framework outlined in our own GEO Guide: A Holistic Blueprint.
We identified 100 prompts prospective clients might use to find agencies like ours — these ranged from “What is the best PR agency for consumer tech?” to “How can we get products featured in AI search?” Our list of prompts has grown to over 260 today as our thinking has evolved, but we started smaller.
We benchmarked Proper Propaganda against both niche consumer tech agencies (like Max Borges Agency and Spark) and holding-company giants (like Edelman and Walker Sands). We don’t typically get into pitches against the bigger holding company agencies, but we decided they probably had the most resources and so were worth including in our competitor comparisons.
We used Scrunch, our favorite AI visibility monitoring platform, to establish our baseline presence across ChatGPT, Google AI Overviews, and Perplexity.
Result: Zero visibility. We didn’t appear in any AI answers. That became the starting point.
3. Fixing the Foundation
Our first major move was a technical and content overhaul of our site. It was long overdue and our house was a mess. Working with our web developer, we restructured pages to align with the S-E-T framework (Structure, Explainability, Trustworthiness).
Key updates included:
- Creating an FAQ hub built from real AI prompt phrasing (“How can PR influence AI search?”, “What is share of scrape?” etc.)
- Applying structured markup and clear H1/H2 hierarchies for crawlability
- Updating metadata and internal links to support AI-friendly entity clustering
- Rewriting existing service and blog pages with short paragraphs, literal explanations, and cited examples
We treated the site like a knowledge graph for bots — readable by humans, but structured for large language models. Bots now make up the majority of internet traffic and, thanks to Scrunch’s bot monitoring features, we know that they exceed human visitors to our site by almost 100x. While we still think our site is pretty and has appeal for humans, the main visitors do not smoke, drink or breathe.
4. Content Expansion and Third-Party Seeding
We realized AI models weigh external authority signals heavily, and that in our industry these were not as plentiful as in others where there are reams of regular media coverage.
To accomplish this we:
- Updated and optimized every blog post using GEO best practices (natural-language Q&A, schema-friendly formatting, quotable statements)
- Launched our Resource Center page, featuring tools, podcasts, case studies, and our full Holistic Blueprint PDF
- Syndicated content to Substack and our Founder’s LinkedIn (especially as articles, since these are more scraped by AI)
- Secured citations on trusted external sources, increasing the number of third-party mentions that AI engines reference when responding to GEO-related prompts
5. Monitoring and Iteration
With Scrunch as our dashboard, we’ve tracked our share of answer across engines weekly and monthly. We used the same methodology we apply to client programs: visibility benchmarking, sentiment review, and where we ranked in terms of position in answers.
6. Results
It was a slow start. We only started to see traction about 14 weeks into the program. However from there things took off like a rocket ship as the gains from AI compounded and as our content production was refined based on data. By November 2025 — four months into the game — we had achieved measurable, defensible gains.
At present, in aggregate, the only two agencies with greater presence across 267 prompts are the giants, Edelman and Walker Sands.
| Metric | Result | Notes |
| Overall AI Share of Answer | 10% | Trailing only Edelman and Walker Sands |
| Perplexity Share of Answer | 15% (category leader) | Ahead of all competitor agencies |
| Prompt Coverage | 2× the volume of large agencies like 5W | Based on tracked prompt universe |
| Presence vs Max Borges | 5× higher visibility | In consumer tech / PR category prompts |
| Inbound Leads via AI Search | 2 signed clients + several active discussions | Directly traceable to AI queries |
| Earned Media Outcomes | 5 podcast appearances for our Founder, including the largest PR podcast globally | Attribution to GEO visibility surge |
7. Takeaways
- Treat GEO as a slow channel. Depending on your industry you won’t see gains quickly. Our first signal of success took 3.5 months.
- Structure beats style. Bots prioritize clear markup, logical page design, and literal phrasing over brand fluff.
- External authority matters more than ever. Media, affiliates, and structured third-party content amplify your AI visibility.
- Measure like a hawk. Use tools like Scrunch to track share-of-answer, citation frequency, and prompt presence continuously.
- Own your category early. Once an AI engine “learns” from you, it seems to keep reinforcing you — compounding over time.
8. Next Steps
Download the full GEO Guide: A Holistic Blueprint and start your own AI-search visibility program.
Or reach out to us at info@properpropaganda.net to discuss your brand’s readiness for AI engines.