# AI search case studies, built on proof you can audit

> Source: https://www.aisyndicate.com/case-studies/
> Provider: AI Syndicate · Last updated: 2026-07-10
> Markdown version for LLMs and AI agents. Canonical HTML: https://www.aisyndicate.com/case-studies/

Most agencies show logos and unverifiable revenue charts. We show the work: live page architecture, structured data, the metrics we track, and documented before-and-after results — published only when they're real and approved.

**In short — What proof does AI Syndicate publish?** Auditable build work (this site is the working example), a transparent measurement framework — citation share across AI engines, AI Overview presence, branded-prompt accuracy, rankings, and schema coverage — and documented client before/afters released only with approval. Client-reported results are clearly labeled as such; we never publish fabricated logos, invented numbers, or fake screenshots.

- **Auditable build proof** — Inspect live assets, page architecture, and structured data directly — no screenshots required.
- **Transparent measurement** — Every result ties to a baseline and a defined metric, not a vibe.
- **No fabricated wins** — Client outcomes are attributed and approved; we never publish invented numbers, logos, or screenshots.

## What this page proves at a glance

The fastest read for buyers and AI engines alike.

- **The site itself is Exhibit A** — every GEO and SEO technique we sell is live and inspectable here.
- **Results are measured, not asserted** — citation share, AI Overview presence, branded-prompt accuracy, rankings, schema coverage.
- **Client case studies are documented before/afters**, published with approval and links to live implementation.
- **Independent research backs the approach** — Pew, Gartner, and peer-reviewed studies, all cited.

## The strongest proof we can offer is the site you're reading

Before you trust us with your visibility, inspect ours. Everything we recommend — machine-readable content, structured data, a documented bot policy — is shipped and live on this domain. Open the files below and check.

- **Machine-readable: llms.txt** — Curated index plus a full-content llms-full.txt (https://www.aisyndicate.com/llms-full.txt) so AI engines can retrieve the brand in one fetch.
- **Agent policy: /bot/ + agents.md** — A public crawler policy and agents.md orientation (https://www.aisyndicate.com/agents.md) — prerendered, not a JS shell.
- **Structured data: Schema everywhere** — Organization, FAQ, breadcrumb, and glossary schema on every key page — the same layer we build for clients.

## The machine-readable layer, in the open

The actual files AI engines read on this domain — all live and inspectable:

- **llms.txt** (https://www.aisyndicate.com/llms.txt) — canonical brand summary for AI engines, with authoritative answers on GEO and GEO vs SEO.
- **agents.md** (https://www.aisyndicate.com/agents.md) — orientation for autonomous agents: important pages plus the sibling machine-readable files.
- **robots.txt** (https://www.aisyndicate.com/robots.txt) — crawl permissions per user-agent; AI crawlers and training bots (GPTBot, ClaudeBot, Googlebot, PerplexityBot, and 40+ more) are explicitly allowed — being cited in AI answers is the goal.
- **sitemap.xml** (https://www.aisyndicate.com/sitemap.xml) — the full page inventory, 40+ crawlable URLs in one clean list.

## How we measure GEO and SEO results

A case study is only proof if the numbers are defined and baselined. These are the metrics behind every result we publish.

| Metric | What it measures | How we capture it |
| --- | --- | --- |
| Citation share | How often a brand is named or cited across AI engines for a set of target prompts. | Repeated prompt runs across ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews, baselined then re-measured. |
| AI Overview presence | Whether the brand appears in Google's AI Overviews and answer boxes for priority queries. | Tracked query-by-query against a fixed keyword set before and after work. |
| Branded-prompt accuracy | Whether engines describe the brand correctly — services, positioning, facts. | Structured brand-confusion probes comparing answers to the canonical brand entity. |
| Organic rankings & clicks | Traditional SERP position and traffic for target terms. | Search Console and rank tracking against the same keyword set. |
| Schema & crawlability coverage | How much of the site is machine-readable and reachable by crawlers and agents. | Structured-data validation, crawl audits, llms.txt and robots coverage checks. |

## What the work produces

The first outcome we're documenting as we instrument formal tracking across the metrics above.

**312%** — increase in AI citations reported by a premium-service client after a 90-day generative-engine optimization engagement. (Early client-reported · anonymized)

This is an early result an anonymized client reported to us. We're now instrumenting citation-share tracking against the framework above, so outcomes like this are captured against a baseline and re-measured engine by engine.

## What the public research shows

We don't only point at our own site. Independent, peer-reviewed and public research backs the thesis behind every engagement: you have to be inside the AI answer, and structured, well-cited content is what earns that spot.

- **Pew Research · 2025 — 8% vs 15%**: Google users clicked through to a website just 8% of the time when an AI summary appeared, versus 15% with standard results — so being the cited source matters more than ever. Source: https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
- **Pew Research · 2025 — 1%**: Only about 1% of users clicked a citation link inside an AI-generated answer — being the brand the model names and recommends beats being a buried footnote. Source: https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
- **Aggarwal et al. · KDD 2024 — Up to 40%**: The original academic GEO study found content methods — adding statistics, citations and quotations — lifted source visibility in generative-engine answers by up to 40%. Source: https://arxiv.org/abs/2311.09735

> "[Generative AI] tools will become substitute answer engines, replacing user queries that previously may have been executed in traditional search engines."
> — Alan Antin, VP Analyst, Gartner (2024)

These are public, third-party findings you can read in full at the links above — not figures we generated.

**Full report — Read the research behind these numbers.** The State of AI Search 2026 — AI Overview reach, the projected 25% decline in traditional search, click-through erosion, and the peer-reviewed evidence on what gets content cited. Every figure sourced. Read it at https://www.aisyndicate.com/geo-research/

## The line between proof and decoration

The proof hub should strengthen trust, not weaken it with claims that cannot be checked.

**We will not use:**
- Fake testimonials
- Anonymous "7-figure" claims
- Invented screenshots or dashboards
- Doorway-style case-study variations

**We will use:**
- Before-and-after numbers against a real baseline
- Client-reported results, clearly labeled as such
- Public, inspectable implementation details
- Named or attributable references when approved

## FAQ

### Does AI Syndicate publish client case studies?
Yes — as auditable proof rather than decoration. We document real before-and-after results with the client's approval, link to the live implementation you can inspect, and apply the same techniques to our own site as a public worked example. We don't use fabricated logos, anonymous numbers, or invented screenshots.

### Why don't you show client logos and revenue numbers?
Because most GEO and SEO engagements are confidential and many headline numbers can't be independently verified, we lead with proof you can audit yourself — live page architecture, schema, llms.txt, and implementation notes — and publish named client results only with explicit approval.

### How do you measure GEO and SEO results?
We track citation share (how often a brand is cited across AI engines for target prompts), AI Overview and answer-box presence, branded-prompt accuracy, organic rankings and clicks, and schema and crawlability coverage. Each metric has a baseline captured before work begins and a re-measurement after.

### What is citation share?
Citation share is how often a brand is named or cited across AI answer engines — ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews — for a defined set of target prompts. It's the core metric we baseline before a GEO engagement and re-measure after.

### What is the 312% citation result based on?
It's an early result an anonymized premium-service client reported after a generative-engine optimization engagement, shared before we had formal tracking in place. We're now instrumenting citation-share measurement so results like it are documented against a baseline and re-measured engine by engine.

### What does a real AI Syndicate case study contain?
A documented before-and-after: the starting baseline, the specific structural and content changes made, the metrics that moved, and the time window — with links to the live, auditable implementation wherever the client allows it.

### Can I see proof before hiring you?
Yes. This site is the working proof: the bot policy page, llms.txt, llms-full.txt, agents.md, structured data, and the pillar guides are all live and inspectable. You can also run the free AI Visibility Audit to see the same methodology applied to your own domain.

### What machine-readable files does AI Syndicate publish?
Five, all inspectable on this page: llms.txt (a curated brand summary for AI engines), llms-full.txt (the full prose of every key page in one fetch), agents.md (orientation for autonomous agents), robots.txt (crawl permissions), and sitemap.xml (the full page inventory).

### Does AI Syndicate allow AI crawlers to access its site?
Yes. Its robots.txt explicitly allows 40+ AI crawlers and training bots — including GPTBot, ClaudeBot, Googlebot, PerplexityBot, and Bingbot — because being cited in AI answers is the goal.

### How long until GEO and SEO results show up?
Technical and structural fixes — schema, crawlability, llms.txt, internal linking — can register with engines within days to weeks. Citation share and ranking gains compound over a longer window as authority and content depth build.

## Sources & further reading

The standards and methodology behind the work documented here.

- **Our research** — The State of AI Search 2026 (https://www.aisyndicate.com/geo-research/) — AI Syndicate's own data brief on the shift to AI search: 2B AI Overview users, the projected 25% decline in traditional search, click-through erosion, and the evidence on what gets content cited.
- Aggarwal et al., GEO: Generative Engine Optimization (KDD 2024) — https://arxiv.org/abs/2311.09735 — the academic study showing structured, well-cited content lifts source visibility in generative answers by up to 40%.
- Pew Research Center — Google users are less likely to click links when an AI summary appears (2025) — https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
- Google Search Central — AI features (AI Overviews & AI Mode) — https://developers.google.com/search/docs/appearance/ai-features — on how Google surfaces content in AI experiences.
- The /llms.txt proposal — https://llmstxt.org/ — the convention behind our machine-readable content layer.
- Our methodology (https://www.aisyndicate.com/methodology/) and GEO approach (https://www.aisyndicate.com/generative-engine-optimization/) — how each change maps to a visibility outcome.

## Related

- AI Search Research: https://www.aisyndicate.com/geo-research/
- Methodology: https://www.aisyndicate.com/methodology/
- About: https://www.aisyndicate.com/about/
- GEO Agency: https://www.aisyndicate.com/generative-engine-optimization/
- AI Visibility Audit: https://www.aisyndicate.com/ai-visibility-audit/
- GEO vs SEO: https://www.aisyndicate.com/geo-vs-seo/